Systems, methods, and computer programs for imaging an object and generating a measure of authenticity of the object

Abstract

An imaging system (200) for imaging and generating a measure of authenticity of an object (10) comprises a dispersive imaging arrangement (30) and an image sensor arrangement (60). They are positioned so that, when electromagnetic radiation (20) from the object (10) illuminates the dispersive imaging arrangement (30), the radiation splits out in different directions into at least a non-dispersed part (40) and a dispersed part (50), and those are imaged by the image sensor arrangement (60). The imaging system (200) is configured to then generate a measure of authenticity of the object (10) depending at least on a relation between the imaged dispersed part, the imaged non-dispersed part, and reference spectral information. The invention also relates to imaging methods, computer programs, computer program products, and storage mediums.

Claims

1. An imaging system for imaging an object and generating a measure of authenticity of the object, the imaging system comprising: an image sensor arrangement having one or more image sensors; and a dispersive imaging arrangement having one or more optical elements, wherein the dispersive imaging arrangement is so that, when electromagnetic radiation from the object illuminates the dispersive imaging arrangement, at least part of the electromagnetic radiation splits out in different directions into at least a non-dispersed part and a dispersed part; and positioned relative to the image sensor arrangement in such a manner as to allow the image sensor arrangement to image said non-dispersed part in a first portion of the image sensor arrangement, so as to obtain a non-dispersed image, and said dispersed part in a second portion of the image sensor arrangement, so as to obtain a dispersed image; the imaging system being configured for, after the image sensor arrangement has, in at least one imaging period, imaged the non-dispersed part and the dispersed part, generating a measure of authenticity of the object depending at least on a relation between the imaged dispersed part, the imaged non-dispersed part, and reference spectral information, wherein the generating of the measure of authenticity comprises one of: computing a synthetic image by deconvolving the imaged dispersed part by the imaged non-dispersed part, and determining the extent to which the synthetic image corresponds to the reference spectral information; computing a synthetic image by deconvolving the imaged dispersed part by the reference spectral information, and determining the extent to which the synthetic image corresponds to the imaged non-dispersed part; and computing a synthetic image by convolving the imaged non-dispersed part and the reference spectral information, and determining the extent to which the synthetic image corresponds to the imaged dispersed part.

2. The imaging system according to claim 1, wherein the dispersive imaging arrangement comprises at least one of: a diffractive element, a transmission diffraction grating, a blazed transmission diffraction grating, a volume holographic grating, a reflective diffraction grating, an arrangement comprising a beam splitter and a diffraction grating, and an arrangement comprising a beam splitter and a dispersive prism.

3. The imaging system according to claim 1, for imaging an object bearing a marking.

4. The imaging system of claim 3, wherein generating the measure of authenticity further comprises decoding a code from the marking within the imaged non-dispersed part and verifying the authenticity of the code.

5. The imaging system of claim 3, wherein the marking comprises at least one machine readable code.

6. An imaging method for imaging an object and generating a measure of authenticity of the object, the imaging method making use of: an image sensor arrangement having one or more image sensors; and a dispersive imaging arrangement having one or more optical elements, wherein the dispersive imaging arrangement is so that, when electromagnetic radiation from the object illuminates the dispersive imaging arrangement, at least part of the electromagnetic radiation splits out in different directions into at least a non-dispersed part and a dispersed part; and positioned relative to the image sensor arrangement in such a manner as to allow the image sensor arrangement to image said non-dispersed part in a first portion of the image sensor arrangement, so as to obtain a non-dispersed image, and said dispersed part in a second portion of the image sensor arrangement, so as to obtain a non-dispersed image; and the imaging method comprising: imaging, by the image sensor arrangement, in at least one imaging period, the non-dispersed part and the dispersed part, and generating a measure of authenticity of the object depending at least on a relation between the imaged dispersed part, the imaged non-dispersed part, and reference spectral information, wherein the generating of the measure of authenticity comprises one of: computing a synthetic image by deconvolving the imaged dispersed part by the imaged non-dispersed part, and determining the extent to which the synthetic image corresponds to the reference spectral information; computing a synthetic image by deconvolving the imaged dispersed part by the reference spectral information, and determining the extent to which the synthetic image corresponds to the imaged non-dispersed part; and computing a synthetic image by convolving the imaged non-dispersed part and the reference spectral information, and determining the extent to which the synthetic image corresponds to the imaged dispersed part.

7. The imaging method according to claim 6, comprising imaging, by the image sensor arrangement, in a plurality of illumination periods, the non-dispersed part and the dispersed part, wherein generating the measure of authenticity comprises: generating, for each illumination period, an intermediate measure of authenticity depending at least on a relation between the dispersed part imaged at the illumination period, the imaged non-dispersed part at the illumination period, and a part of the reference spectral information, said part of the reference spectral information being associated with how the object has been illuminated during the illumination period; and generating the measure of authenticity based on the plurality of generated intermediate measures of authenticity.

8. The imaging method according to claim 6, comprising imaging, by the image sensor arrangement, in a plurality of illumination periods, the non-dispersed part and the dispersed part, wherein generating the measure of authenticity comprises: processing the imaged non-dispersed part based at least on the non-dispersed part imaged at a first illumination period among the plurality of illumination periods and the non-dispersed part imaged at a second illumination period among the plurality of illumination periods, wherein the illumination conditions during the first illumination period at least partially differ from the illumination conditions during the second illumination period; processing the imaged dispersed part based at least on the dispersed part imaged at the first illumination period and the dispersed part imaged at the second illumination period; and generating the measure of authenticity depending at least on a relation between the processed imaged dispersed part, the processed imaged non-dispersed part, and the reference spectral information.

9. The imaging method according to claim 6, for imaging an object bearing a marking.

10. The imaging method of claim 9, wherein generating the measure of authenticity further comprises decoding a code from the marking within the imaged non-dispersed part and verifying the authenticity of the code.

11. The imaging method according to claim 6, further comprising a step of controlled illumination of the object.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Embodiments of the present invention shall now be described, in conjunction with the appended figures, in which:

(2) FIG. 1 schematically illustrates an object to be imaged and an imaging system in one embodiment of the invention;

(3) FIG. 2 schematically illustrates an object to be imaged and a system in one embodiment of the invention, wherein the system comprises both an imaging system and an illumination arrangement;

(4) FIG. 3 schematically illustrates an object to be imaged and a system in one embodiment of the invention, wherein the system notably comprises illumination elements arranged around a dispersive imaging arrangement;

(5) FIGS. 4 to 6 schematically illustrate three imaging systems and objects to be imaged, in three embodiments of the invention respectively;

(6) FIGS. 7 and 8 schematically represent, using a thin lens-grating approximation, two imaging systems and marks to be imaged, in two embodiments of the invention respectively, wherein FIG. 8 especially illustrates the order separation;

(7) FIG. 9a schematically illustrates an imaging system in one embodiment of the invention, wherein the imaging system is an imaging device;

(8) FIG. 9b schematically illustrates a system in one embodiment of the invention, wherein the system comprises both an imaging system and an illumination arrangement, and wherein the system is an imaging device;

(9) FIG. 10a schematically illustrates an imaging system in one embodiment of the invention, wherein the imaging system comprises an imaging device and said imaging device comprises an image sensor arrangement and a dispersive imaging arrangement, but said imaging device is not configured to actually generate the measure of authenticity;

(10) FIG. 10b schematically illustrates a system in one embodiment of the invention, wherein the system comprises an imaging device and said imaging device comprises an image sensor arrangement, a dispersive imaging arrangement and an illumination arrangement, but said imaging device is not configured to actually generate the measure of authenticity;

(11) FIG. 11 is a flowchart of an imaging method in one embodiment of the invention;

(12) FIGS. 12a to 12c are flowcharts of imaging methods in three embodiments of the invention, wherein generating the measure of authenticity depends at least on the extent to which the imaged dispersed part corresponds to a convolution of the imaged non-dispersed part and reference spectral information;

(13) FIG. 13 is a flowchart of an imaging method in one embodiment of the invention, involving the decoding of a code from a marking within the imaged non-dispersed part and verifying the authenticity of the code;

(14) FIG. 14a schematically illustrates an imaging system in one embodiment of the invention, when applied, by simulation, to a single dot of a two-dimensional matrix barcode;

(15) FIG. 14b schematically illustrates an imaging system in one embodiment of the invention, when applied, by simulation, to a two-dimensional matrix barcode;

(16) FIG. 15 shows exemplary zero- and first-order images of two-dimensional matrix barcodes printed on labels, imaged by an imaging system in one embodiment of the invention;

(17) FIG. 16 shows the exemplary result (upper-right chart) of the column-by-column deconvolution (or similar non-linear process) from a two-dimensional matrix barcode image containing zero- and first-order components (left-hand image), as well as the comparison of the average of all spectrum curves of the upper-right chart to reference spectral information (lower-right chart);

(18) FIGS. 17 to 19 schematically illustrate three imaging systems in three embodiments of the invention, respectively;

(19) FIGS. 20 and 21 schematically illustrate the generation of a measure of authenticity of an object, in two embodiments of the invention, wherein the image sensor arrangement images the non-dispersed part and dispersed part in a plurality of illumination periods;

(20) FIGS. 22, 23a and 23b are flowcharts of imaging methods in three embodiments of the invention, wherein the generation of the measure of authenticity of an object follows the image sensor arrangement imaging the non-dispersed part and dispersed part in a plurality of illumination periods;

(21) FIGS. 24a and 24b show images of a soft-drink can cap without (FIG. 24a) and with a mask (FIG. 24b) acquired using an imaging system in one embodiment of the invention;

(22) FIG. 25 shows examples of images of a soft-drink can cap acquired without a physical mask but excited in two different illumination periods by blue light (left-hand image) and green light (right-hand image), in one embodiment of the invention;

(23) FIG. 26 shows examples of background-subtracted images using two different linear combinations, in one embodiment of the invention;

(24) FIG. 27 shows examples of extracted spectra with and without applying the DIBS algorithm on acquired images, in one embodiment of the invention;

(25) FIG. 28 shows the spectral reflectivity of two different colour pigments;

(26) FIG. 29 shows the typical relative spectral distribution of a white LED;

(27) FIG. 30 shows the typical relative spectral distribution of an incandescence bulb at 3000 K temperature, compared to the one of the sun;

(28) FIG. 31 shows the excitation spectrum and emission spectrum of an exemplary fluorescent dye;

(29) FIGS. 32 and 33 show the emission and excitation spectra for exemplary phosphorescent phosphor pigments;

(30) FIG. 34 is a schematic diagram of an exemplary implementation of a computing unit according to one embodiment of the invention;

(31) FIGS. 35a to 35d schematically illustrate examples of imaging period(s) and illumination period, in four embodiments of the invention; and

(32) FIG. 36 schematically illustrates an imaging system comprising, on the one hand, an imaging device comprising an image sensor arrangement, wherein the imaging device is a mobile phone having a camera, and, on the other hand, an imaging accessory comprising a dispersive imaging arrangement.

DETAILED DESCRIPTION

(33) The present invention shall now be described in conjunction with specific embodiments. These specific embodiments serve to provide the skilled person with a better understanding, but are not intended to restrict the scope of the invention, which is defined by the appended claims. A list of abbreviations and their meaning is provided at the end of the detailed description.

(34) FIG. 1 schematically illustrates an imaging system 200 in one embodiment of the invention. System 200 aims at imaging an object 10 and generating a measure of authenticity of object 10, i.e. an article. Object 10 may for example be, without being limited to, a bottle or can of beer, wine, liquor or soft-drink, a pack, packet or box of cigarettes or cigars, a medicine pack, a bottle of perfume, or any other excisable goods, a banknote, a value paper, an identity document, a card, ticket, label, banderol, security foil, security thread or the like. Object 10 has at least one part, surface or side bearing a visible or invisible mark, logo, sign, image, or pattern, for example printed with a printing ink and/or coating, either printed on a label apposed on object 10 or printed directly on object 10 (such as on a cap, capsule or the like of object 10, wherein the cap or capsule may for example have a coloured background). The expected spectral response of said part, surface or side, and possibly the ink thereon (which may or may not have, for example, photoluminescent properties), when subject to particular illumination conditions, is known and constitutes the reference spectral information.

(35) System 200 comprises an arrangement 60, hereinafter referred to as image sensor arrangement 60, consisting in one or more image sensors. System 200 also comprises another arrangement 30, hereinafter referred to as dispersive imaging arrangement 30, consisting in one or more optical elements.

(36) In one embodiment, image sensor arrangement 60 comprises one or more array CCD or CMOS detectors to record the intensity distribution of the incident electromagnetic energy. Dispersive imaging arrangement 30 not only disperses electromagnetic energy but may also gather electromagnetic energy from object 10 and focus the electromagnetic energy rays to produce an image of object 10 onto an image plane where image sensor arrangement 60 is positioned. In one embodiment, dispersive imaging arrangement 30 comprises, on the one hand, at least one of a diffractive element, a refractive element, one or more lenses, and an objective, in order to produce an image of object 10 onto the image plane where image sensor arrangement 60 is positioned, and, on the other hand, a long pass filter (also called long-wavelength pass filter) in order to limit the spectral range used for authentication.

(37) System 200 may also comprise optionally various auxiliary elements (not shown in FIG. 1) such as for example any one or any combination of: a) a housing for containing, covering and/or protecting dispersive imaging arrangement 30 and image sensor arrangement 60; b) supporting elements integrally formed within the housing, or attached thereto, to maintain dispersive imaging arrangement 30 in a fixed or substantially fixed relative position with respect to image sensor arrangement 60; c) a protective cover or protective covering means to be used between object 10 and dispersive imaging arrangement 30 to avoid parasitic illumination from ambient light and/or sunlight (in this case, a controlled illumination source may be contained within this protective cover); d) additional optical filters (long-pass, bandpass, etc.), which may for example be advantageous if imaging system 200 operates in luminescence mode, to cut out the irradiation source reflection; e) a controller or controlling means or units for controlling the operation of image sensor arrangement 60 and other elements; f) outputting and inputting means for providing information to and receiving information from an operator, such as a display screen, a keyboard, push-buttons, control knobs, LED indicator lights, etc. (in that respect, see also FIG. 34 and the corresponding description); and g) a battery for powering various electronic parts of system 200.

(38) Dispersive imaging arrangement 30 is constituted and positioned so that, when electromagnetic radiation 20 from object 10 illuminates arrangement 30 or in particular a specific part, surface, side, aperture, or opening thereof, at least part of radiation 20 splits out in different directions into at least a non-dispersed part 40 and a dispersed part 50. The word dispersive means here: that separates in its constituent wavelength components. Arrangement 30 may for example comprise: a diffractive element, a transmission diffraction grating (also known simply as transmission grating, or rarely as transmissive diffraction grating), a blazed transmission diffraction grating, a volume holographic grating, a grism (also called grating prism), a reflective diffraction grating, an arrangement comprising a beam splitter and a diffraction grating, an arrangement comprising a beam splitter and a dispersive prism, or a combination of any of those. If arrangement 30 diffracts radiation 20, non-dispersed part 40 may be referred to as the zero diffraction order part of the radiation, and dispersed part 50 may be referred to as a non-zero diffraction order part, such as for example the negative or positive first diffraction order part of the radiation.

(39) Here are some examples of transmission gratings that may be used in some embodiments of the invention: Example 1: Especially for a transmission grating mounted in front of an objective (see also in that respect FIGS. 4 and 17), a Thorlabs # GT13-06V (from Thorlabs, Inc., based in Newton, N.J., U.S.) with grooves density 600 lines per mm (I/mm), blaze angle 28.7, size 12.712.73 mm from Schott B270 glass, may be used. Example 2: Especially for a transmission grating mounted between an objective and the image sensor(s) (see also in that respect FIGS. 5, 6, 18 and 19), a Richardson grating 340056TB07-775R (from Newport Corporation, based in Rochester, N.Y., U.S.) with grooves density of 360 I/mm, blaze angle 21, and size 12.712.73 mm, may be used. Example 3: Especially for a back-mounted grating for extended field of view, a Thorlabs # GTU13-06 with grooves density 600 I/mm, blaze angle 22, and size 12.712.72 mm from fused silica, may be used.

(40) Electromagnetic radiation 20 coming from object 10 and illuminating dispersive imaging arrangement 30 may originate in part or in full from the reflection of electromagnetic radiation emitted by an electromagnetic radiation source (not shown in FIG. 1). Radiation 20 from object 10 and illuminating arrangement 30 may alternatively, or additionally, originate in part or in full from some form of photoluminescence (i.e., fluorescence or phosphorescence) of a substance of object 10 upon or after the illumination of object 10 by electromagnetic radiation emitted by an electromagnetic radiation source. In both cases (i.e., radiation by reflection or by some form of photoluminescence), the electromagnetic radiation source may, in one embodiment, be integrated with, or attached to, a housing containing imaging system 200 (or part thereof). Said electromagnetic radiation source may for example be a light source, an infrared radiating source, and/or an UV radiating source. In one embodiment, the electromagnetic radiation source is an illumination source controlled by, or together with, system 200.

(41) Electromagnetic radiation 20 coming from object 10 usually contains radiation of more than one wavelength, especially when object 10 is authentic. That is, radiation 20 is usually polychromatic in the broad sense of the term, i.e. not necessarily limited to visible colours. Radiation 20 may for example be in any wavelength range encompassed between 180 nm (UV radiation) and 2500 nm (infrared radiation), i.e. in the visible light range and/or outside that range (for example in the near-infrared (NIR) or short-wavelength infrared (SWIR) range). The portion of radiation 20 reaching dispersive imaging arrangement 30 that is actually dispersed may depend on the characteristics of the optical element(s) forming arrangement 30. For example, long pass filter may be used to select the spectral range to be analysed.

(42) Furthermore, dispersive imaging arrangement 30 is positioned relative to image sensor arrangement 60 in such a manner as to allow arrangement 60 to simultaneously image in one imaging period (as illustrated by FIG. 35a), to sequentially image in two imaging periods (as illustrated by FIG. 35b), or to partially simultaneously image in two imaging periods (as illustrated by FIGS. 35c and 35d), non-dispersed part 40 in a first portion of arrangement 60 and dispersed part 50 in a second portion of arrangement 60.

(43) An example of image sensor that may be used in some embodiments of the invention is: a 1/3-Inch Wide-VGA CMOS Digital Image Sensor MT9V022 from ON Semiconductor, based in Phoenix, Ariz., U.S. That sensor has 752-by-480 pixels with size 6 m forming active imager size with dimensions of 4.51 mm2.88 mm and diagonal of 5.35 mm.

(44) An imaging period is here defined as being: a) if non-dispersed part 40 and dispersed part 50 are simultaneously acquired by image sensor arrangement 60, the period during which both non-dispersed part 40 and dispersed part 50 are acquired (as illustrated by FIG. 35a), or b) if non-dispersed part 40 and dispersed part 50 are sequentially (as illustrated by FIG. 35b) or partially simultaneously (as illustrated by FIGS. 35c and 35d) acquired by image sensor arrangement 60, each of the period during which non-dispersed part 40 is acquired and the period during which dispersed part 50 is acquired.

(45) In one embodiment, each or at least one imaging period has a duration having a value selected from the range of 5 to 1200 ms, and preferably selected from the range of 10 to 800 ms, such as for example 10, 20, 30, 50, 75, 100, 150, 200, or 300 ms.

(46) In one embodiment, the duration of the imaging period for imaging non-dispersed part 40 and the duration of the imaging period for imaging dispersed part 50 differ from each other. This embodiment may be advantageous in particular when using diffraction gratings having different efficiencies for the zero- and first-order. For example, the duration of the imaging period for imaging non-dispersed part 40 may be 10 ms, whereas the duration of the imaging period for imaging dispersed part 50 may be 100 ms.

(47) An illumination period (as illustrated by FIGS. 35a to 35d) is here defined as being a period during which illumination conditions are considered sufficiently constant for the purpose of imaging non-dispersed part 40 and dispersed part 50, and generating a measure of authenticity based thereof.

(48) In one embodiment, the first and second portions of image sensor arrangement 60 are on two different image sensors of arrangement 60. When using two image sensors for imaging non-dispersed and dispersed parts 40, 50 their relative positioning has to be taken into account.

(49) In another embodiment, the first and second portions of arrangement 60 are two different portions of a single image sensor. In other words, in this embodiment, non-dispersed and dispersed parts 40, 50 may be captured in a single frame.

(50) The configuration (geometry, parameters, etc.) of the optical elements of dispersive imaging arrangement 60 allows the separation of dispersed part 50 from non-dispersed part 40 from within the single frame. Shorter wavelengths are less deflected than longer wavelengths. In one embodiment, system 200 is configured to avoid overlapping of first-order image at the shortest wavelength with the zero-order image (see also FIG. 8, which schematically illustrates the order separation). A long pass filter may, for example, be used to cut shorter wavelengths as for example shown in FIG. 8, in order to prevent overlapping of orders.

(51) The portion of electromagnetic radiation 20 illuminating and passing through dispersive imaging arrangement 30 (therefore being dispersed in one set of directions and being non-dispersed in another set of directions) that is then actually detected by image sensor arrangement 60 depends on the characteristics of its image sensor(s). The electromagnetic radiation detected by the image sensor(s) may for example be in any wavelength range encompassed between 180 nm (UV radiation) and 2500 nm (infrared radiation), i.e. in the visible light range and/or outside that range (for example in the near-infrared (NIR) or short-wavelength infrared (SWIR) range). In that example, the lower limit of 180 nm may be imposed by material constraints of both dispersive imaging arrangement 30 and image sensor(s) 60, whereas the upper limit of 2500 nm may for example be imposed by the spectral response of indium gallium arsenide-based (GalnAs) infrared detectors. In one embodiment, the electromagnetic radiation detected by image sensor(s) 60 is in the range of visible light. In one embodiment, the electromagnetic radiation detected by image sensor(s) 60 is in the wavelength range of 180 nm to 2500 nm, more preferably in the range of 400 nm to 1000 nm.

(52) Yet furthermore, imaging system 200 is configured for, after image sensor arrangement 60 has imaged non-dispersed part 40 and dispersed part 50 in at least one imaging period, generating a measure of authenticity of object 10 depending at least on a relation between the imaged dispersed part, the imaged non-dispersed part, and reference spectral information. System 200 thus enables the verification of whether, and/or the extent to which, the relation between the imaged dispersed part, the imaged non-dispersed part, and the reference spectral information, which represents the expected spectral composition of electromagnetic radiation 20 coming from object 10, is in accordance with the expected underlying physics of the system. If so, object 10 is likely to be authentic. Otherwise, it is more likely to be a counterfeit. System 200 thus enables a form of material-based authentication, such as for example at least one of: a) material-based authentication of the ink used to create a mark printed on object 10, and b) material-based authentication of object 10 itself especially if object 10 is luminescing with a specific emission spectrum or has a specific reflection or absorption spectrum.

(53) The nature of the relation that is looked at, i.e. the relation between the imaged dispersed part, the imaged non-dispersed part, and the reference spectral information, may be understood in the sense that, if the reference spectral information corresponds or substantially corresponds to the spectral composition of electromagnetic radiation 20 coming from imaged object 10, the imaged dispersed part typically resembles (non-linear effects may also need to be taken into account) the result of the convolution of the imaged non-dispersed part with the reference spectral information, in which case object 10 is likely to be authentic. In contrast, if the reference spectral information does not correspond to the spectral composition of radiation 20 coming from imaged object 10, the imaged dispersed part typically noticeably differs from the result of the convolution of the imaged non-dispersed part with the reference spectral information, in which case object 10 is likely to be a counterfeit.

(54) More generally, the nature of the relation that is looked at, i.e. the relation between the imaged dispersed part, the imaged non-dispersed part, and the reference spectral information, may also significantly differ from a mere convolution, considering the existence of non-linear effects. The nature of the relation may be determined a) based on the underlying physics and geometry, b) empirically, and/or c) by simulation (for example, using raytracing methods of commercially available solutions, such as e.g. Zemax optical design program, available from Zemax, LLC, based in Redmond, Wash., U.S.).

(55) The underlying physics and geometry may include (i) the properties of dispersive imaging arrangement 30, image sensor arrangement 60, the transmission medium in between, etc., and (ii) effects of stretch of the image (zero- or first-order) in the direction of the dispersion (y axis), which may be compensated for by mapping of the y axis of the image (zero- or first-order) to a new y axis using a non-linear function. The image may be stretched due to 1) non-linear dispersion of the grating, 2) projection distortions (with different paths from arrangement 30 to arrangement 60), and/or 3) optics-specific field aberrations (as lenses may distort slightly differently the zero- and first-order).

(56) The non-linear effects may also, in one embodiment, be modelled as a relation between the dispersed and non-dispersed images and the reference spectrum in a form being as close to linear translation-invariant (LTI) as possible. In such a case, the determination of the non-linear effects may be performed for example by a) acquiring several zero- and first-order images of objects 10 with a known reference spectrum, and b) fitting the non-linear parameters to transform the relation to LTI.

(57) One way to determine the non-linear effects, and therefore the nature of the relation to be looked at, may be a mathematical analysis of the optical system and determination of the correction that has to or should be applied to make the system LTI. This may be done using optical equations found for example in textbooks such as Yakov G. Sosking, Field Guide to Diffractive Optics, SPIE, 2011. This may also be done numerically using optical software such as for example Zemax OpticStudio, available from Zemax, LLC.

(58) In one embodiment, dispersive imaging arrangement 30 diffracts electromagnetic radiation 20 using a diffraction grating, and the imaged non-dispersed part is the image in zero diffraction order of the diffraction grating, whereas the imaged dispersed part is the image in a first diffraction order of the diffraction grating. An average spectral irradiance of a region of the image may be reconstructed using the imaged non-dispersed and dispersed parts, and then the average spectral irradiance may be compared to the expected spectral irradiance (the reference spectral information). In one embodiment, the grooves profiles of the diffraction grating (e.g. blaze angle) are optimized to spread most of the input electromagnetic radiation into these two orders.

(59) In one embodiment, generating a measure of authenticity of object 10 comprises authenticating it, i.e. determining that it is likely to be authentic or not. In one embodiment, generating a measure of authenticity of object 10 comprises generating an authenticity measure (or index) such as for example a real value between 0 and 1, wherein 0 may mean fully sure that the object is not authentic and 1 may mean fully sure that the object is authentic.

(60) In practice, the authentication index typically does not reach the value 1 for all authentic objects (and 0 for all non-authentic ones). Hence, in one embodiment, a threshold between 0 and 1 is defined (for example a value comprised between 0.80 and 0.90, and in particular 0.85) above which the object is considered as authentic, and below which the object is considered as non-authentic. This threshold may for example be defined through measurements on a set of authentic and non-authentic objects. These measurements typically produce a bi-modal distribution of indexes (i.e., one part for the authentic objects concentrated towards the value 1 and one part for the non-authentic ones below, both separated by a gap). The robustness of the method is directly related to the extent to which the two parts (modes) of the index distribution are distant from one another. The threshold may then be placed in between either close to the index distribution of the authentic objects to minimize false positives or closer to the non-authentic index distribution to minimize false negatives.

(61) If object 10 is, for example, a container or package containing some goods, the generated measure of authenticity may merely amount to a measure of authenticity of the goods determined through a mark or sign existing on the container or package (assuming that the container or package has not been tampered with), not necessarily directly enabling to authenticate the goods as such.

(62) Since non-dispersed and dispersed parts 40, 50 of the electromagnetic radiation may be imaged in one imaging period, and since the imaging enables the determination of the spectral composition of incident electromagnetic radiation 20, imaging system 200 may be regarded as a form of snapshot spectral imager in the sense that the scene is not scanned during the imaging process. However, system 200 does not enable or at least does not necessarily enable obtaining the spectral composition, i.e. irradiance, of each point (x, y) of the scene, which is as such not necessary for authentication provided that there is a dominant spectral response in the image.

(63) FIG. 2 schematically illustrates an object 10 to be imaged and a system 220 in one embodiment of the invention. System 220 comprises both an imaging system 200 (as described above with reference to FIG. 1) and an illumination arrangement 210. In one embodiment, system 220 forms a single device, such as for example a handheld, code reading and authentication device.

(64) Illumination arrangement 210 generates electromagnetic radiation 21 for illuminating object 10. In one embodiment, radiation 21 has known parameters (e.g., spectrum, power, homogeneity, etc.) to allow excitation of e.g. luminescence emission spectra to allow imaging of object 10 and/or mark thereon and analysing the emission spectra for authentication. As explained above with reference to FIG. 1, electromagnetic radiation 20 originates from object 10, and/or mark thereon, and reaches imaging system 200.

(65) In one embodiment, system 220 is connected to driving electronics and sensor reading electronics, so that, for example, image data outputted by imaging system 200 may be transferred to a processing unit for data treatment.

(66) FIG. 3 schematically illustrates an object 10 to be imaged and a system 220 in one embodiment of the invention, as a possible implementation of the system illustrated on FIG. 2. System 220 notably comprises illumination elements 22 arranged around dispersive imaging arrangement 30. Although two illumination elements 22 are shown in FIG. 3, any number of illumination elements 22 may be provided, such as for example three, four or more. Furthermore, in one embodiment, illumination elements 22 are arranged symmetrically around dispersive imaging arrangement 30. The symmetric arrangement of illumination elements 22 around arrangement 30 is advantageous for homogeneous illumination of the target surface of object 10.

(67) FIGS. 4 to 6 schematically illustrate three imaging systems 200 in three embodiments of the invention, respectively, showing possible components of dispersive imaging arrangement 30, such as a transmission grating 31, an imaging lens 32, an optical long-pass filter 33, and an additional lens arrangement 34.

(68) Arrangement 30 of FIG. 4 comprises an imaging lens 32, a transmission grating 31 mounted in front of lens 32, and an optical long-pass filter 33 mounted behind lens 32. This enables to produce low optical aberrations for both dispersed and non-dispersed images by using the broad field-of-view of the lens objective.

(69) In arrangement 30 of FIG. 5, both transmission grating 31 and optical long-pass filter 33 are mounted behind lens 32. This enables to cancel the dependence of the extracted spectra on the object position along the optical axis.

(70) In the embodiment of FIG. 6, optical long-pass filter 33 is mounted in front of lens 32, and transmission grating 31 is mounted behind lens 32. Furthermore, an additional lens arrangement 34 is also mounted behind lens 32. This configuration enables to efficiently separate the dispersed and non-dispersed images and avoid dependence on the object position along the optical axis.

(71) FIGS. 7 and 8 schematically represent, using a thin lens-grating approximation, two imaging systems 200 and marks 11 in two embodiments of the invention, respectively, to help understand the order separation and definition of minimum wavelength of a spectral range which is analysed to authenticate mark 11.

(72) In FIG. 7, dispersive imaging arrangement 30 includes a lens, a transmission grating and a long-wavelength pass filter, to create non-dispersed image 41 (zero-order) and the dispersed image on the image plane 65 where the image sensor(s) are positioned. Dispersed beams 50-1 are for the shortest wavelength .sub.1 and create dispersed image 51 corresponding to wavelength .sub.1.

(73) Imaging system 200 receives electromagnetic energy 20 originating from object 10 to create a non-dispersed image 41 of object 10 onto image plane 65. Non-dispersed part 40 is produced by arrangement 30 in the same or similar way as an ordinary, non-dispersive imaging arrangement consisting merely of a lens.

(74) The dispersed part is shifted compared to the non-dispersed part and is blurred by the spectrum of electromagnetic energy 20 impinging arrangement 30. The minimum shift depends on the minimum wavelength present in the spectrum emitted by object 10 or depends on the minimum wavelength transmitted through arrangement 30. The minimum shift may also depend on some grating and system parameters (e.g. grooves density, order, and incident angle) which parameters define the angular dispersion of the grating.

(75) The three discrete dispersed images of mark on FIG. 7 correspond to discrete wavelengths .sub.1, .sub.2 and .sub.3. These discrete wavelengths can therefore be conveniently resolved since the corresponding images do not overlap. Furthermore, system 200 separates dispersed image 51 for wavelength .sub.1 from non-dispersed image 41, so that, on the one hand, an image of mark may conveniently be read (e.g., to decode the code represented by the mark) and, on the other hand, the emission spectra of the ink used to print mark may be extracted.

(76) FIG. 8 shows the imaging of an area 12 of object 10, wherein area 12 contains a printed mark 11, which may be in any position or orientation. If mark 10 is outside area 12, imaging system 200 should be repositioned so as to have mark 11 within area 12. Non-dispersed image 41 of area 12 contains the image of mark 11. Dispersed image 51 of area 12 contains the image of mark 11.

(77) Image 51 corresponds to the minimum wavelength .sub.min that can be transmitted by the system and defined by a cut-on wavelength of a long pass filter of arrangement 30. Reference 62 shows the order separation, which, in one embodiment, corresponds to the size of image 41 of area 12. In one embodiment, arrangement 30 enables this order separation for the minimum wavelength .sub.min so as to efficiently authenticate object 10.

(78) In one embodiment, illumination arrangement 210 (not illustrated on FIG. 8) illuminates only the portion of object 10 corresponding to area 12. Illumination arrangement 210, together with an optional protective cover (as mentioned above), may be designed to prevent ambient light from reaching area 12, thus providing better conditions for code reading and authentication.

(79) FIG. 9a schematically illustrates an imaging system 200 in one embodiment of the invention, which differs from imaging system 200 of FIG. 1 in that system 200 of FIG. 9a specifically consists in a single imaging device. In addition to dispersive imaging arrangement 30 and image sensor arrangement 60 described with reference to FIG. 1, system 200 comprises a processing unit 70 configured for receiving data representing the imaged non-dispersed and dispersed parts (as detected by arrangement 60), generating the measure of authenticity as described with reference to FIG. 1, and outputting information 80 representing the generated measure of authenticity to any kind of user interface of the imaging device and/or to an output port for transmission to one or more other external devices (not shown in FIG. 9a).

(80) In one embodiment, the imaging device making up imaging system 200 of FIG. 9a is a hand-held device. Such an imaging device can therefore be regarded as a hand-held audit device capable of generating a measure of authenticity of an object, and providing the authenticity measure to, for example, the device's operator.

(81) FIG. 9b schematically illustrates a system 220 in one embodiment of the invention, wherein system 220 comprises both an imaging system 200 and an illumination arrangement 210, and wherein system 220 is an imaging device. In other words, the embodiment of FIG. 9b may be regarded a combination of the embodiments described with reference to FIGS. 9a and 2. In one embodiment, the imaging device making up system 200 of FIG. 9b is a hand-held device.

(82) FIG. 10a schematically illustrates an imaging system 200 in one embodiment of the invention, which differs from imaging system 200 of FIG. 1 in that system 200 of FIG. 10a is shown as specifically comprising more than one device. Namely, in the example of FIG. 10a, system 200 comprises two devices: on the one hand, an imaging device 100 comprising dispersive imaging arrangement 30 and image sensor arrangement 60 described with reference to FIG. 1, and, on the other hand, a processing device 110 comprising a processing unit 70. Processing device 110, rather than imaging device 100, generates the measure of authenticity (as described with reference to FIG. 1). To do so, data 90 representing the imaged non-dispersed and dispersed parts is transmitted from imaging device 100 to processing device 110. Data 90 may be transmitted on any suitable wired or wireless channel using any transmission format (such as for example using Internet Protocol (IP) packets, optionally encrypted). Then, within processing device 110, the measure of authenticity is generated by processing unit 70, and information 80 representing the generated measure of authenticity may then be outputted to a user interface of processing device 110 and/or to an output port for transmission to one or more other external devices (not shown in FIG. 10a).

(83) FIG. 10b schematically illustrates a system 220 in one embodiment of the invention, wherein system 220 comprises an imaging device 100 and said imaging device 100 comprises an image sensor arrangement 30, a dispersive imaging arrangement 60 and an illumination arrangement 210, but imaging device 100 is not configured to actually generate the measure of authenticity. In other words, the embodiment of FIG. 10b may be regarded a combination of the embodiments described with reference to FIGS. 10a and 2.

(84) In one embodiment, imaging device 100 of any one of FIGS. 10a and 10b is a hand-held device.

(85) In one embodiment, processing unit 70 of any one of FIGS. 9a, 9b, 10a and 10b forms part of a computing unit such as for example the one illustrated with reference to FIG. 34 (which is discussed below). In such a case, processing unit 70 of FIG. 9a or 9b and processing unit 503 of FIG. 34 may actually be the same element. Likewise, in such a case, processing unit 70 of FIG. 10a or 10b (within processing device 110) and processing unit 503 of FIG. 34 may actually be the same element.

(86) In some embodiments, the imaging device making up imaging system 200 of FIG. 9a or 9b, or imaging device 100 illustrated in FIG. 10a or 10b comprises a handle integrally formed with the housing, or attached thereto, to enable an operator to hold the imaging device towards to the object to be imaged and authenticated.

(87) In one embodiment, the imaging device making up imaging system 200 of FIG. 9a or making up system 220 of FIG. 9b, or imaging device 100 illustrated in any one of FIGS. 10a and 10b further comprises a storage unit (not shown in any of FIGS. 9a, 9b, 10a, and 10b) for storing, for example, the reference spectral information which is known in advance and used for generating the measure of authenticity. The reference spectral information may be stored in the form of a reference spectral profile.

(88) FIG. 11 is a flowchart of a method in one embodiment of the invention, which makes use of an image sensor arrangement 60 and a dispersive imaging arrangement 30, as described above with reference to FIGS. 1 to 10b. The method comprises the steps of imaging s300, by arrangement 60, in at least one imaging period, non-dispersed part 40 and dispersed part 50, and generating s400 a measure of authenticity of object 10 depending at least on a relation between the imaged dispersed part, the imaged non-dispersed part, and reference spectral information. Step s400 is carried out through convolution or deconvolution operation(s) (as discussed below with reference to FIGS. 12a to 12c) or through convolution-like or deconvolution-like operation(s) to take into account non-linear effects as explained above.

(89) If imaging step s300 consists in imaging non-dispersed part 40 and dispersed part 50 in a single illumination period, step s300 precedes generating step s400, usually without overlap. However, if step s300 consists in imaging non-dispersed part 40 and dispersed part 50 in a plurality of illumination periods (typically under different illumination conditions), imaging step s300 and generating step s400 may overlap (not shown in FIG. 11). Namely, the process of generating s400 the measure of authenticity may begin based on image data recorded during one or more illumination periods while imaging step s300 is still under way. In one embodiment, generating s400 the measure of authenticity depends at least on the extent to which the imaged dispersed part corresponds to a convolution of the imaged non-dispersed part and the reference spectral information. This may be implemented in different ways as illustrated by FIGS. 12a to 12c.

(90) In particular, in a first sub-embodiment, illustrated by the flowchart of FIG. 12a, generating s400 the measure of authenticity comprises: deconvolving s410 the imaged dispersed part by the imaged non-dispersed part, and determining s420 the extent to which the result corresponds to the reference spectral information.

(91) In a second sub-embodiment, illustrated by the flowchart of FIG. 12b, generating s400 the measure of authenticity comprises: deconvolving s430 the imaged dispersed part by the reference spectral information, and determining s440 the extent to which the result corresponds to the imaged non-dispersed part; and

(92) In a third sub-embodiment, illustrated by the flowchart of FIG. 12c, generating s400 the measure of authenticity comprises: convolving s450 the imaged non-dispersed part and the reference spectral information, and determining s460 the extent to which the result corresponds to the imaged dispersed part.

(93) A possible implementation of step s400 in this third sub-embodiment may be described as follows:

(94) In step s450, a synthetic first diffraction order image is computed by convolving the known spectral signature of the authentic ink (i.e., the reference spectral information) with the zero-order image (i.e., the imaged non-dispersed part). Then, in step s460, a cross-correlation between the acquired first-order image (i.e., the imaged dispersed part) and the synthetic first-order image (i.e., the result of step s450) is used to compare them and generate a similarity parameter. This correlation may be performed not only on the images but also the first and second derivatives of the images to output three similarity parameters. Then, a decision is made by for example applying classifiers based on machine learning algorithms on the similarity parameter sets to authenticate mark on object 10.

(95) A convolution might, however, not always lead to the best results due to the existence of non-linear effects (as discussed above). Thus, in one embodiment of the invention, rather than carrying out a convolution in step s450, a model or function may be used, which may be determined in advance using instrument calibration data, measurements, modelling or a combination thereof. The model or function is a computation model for computing a synthetic first-order image (i.e., a synthetic dispersed part) from a given zero-order image (i.e., the imaged non-dispersed part) and a known spectrum (i.e., the reference spectral information). Similar considerations apply to deconvolving steps s410 and s430, which may be replaced by other models or functions.

(96) In order to carry out the comparison part of step s460 in this implementation, the acquired first-order image (i.e., the imaged dispersed part) and the synthetic first-order image (i.e., the output of step s450) are compared, and one or several matching similarity values are computed.

(97) In one embodiment, the matching value is the cross-correlation value of the two images, i.e. the acquired first-order image and synthetic first-order image. In another embodiment, the matching value is the cross-correlation value of the derivative of the two images. In a further embodiment, the matching value is the cross-correlation value of the second derivative of the two images. In yet another embodiment, more than one matching values are extracted from a combination of the previously proposed matching values. The computations may take place on the entire first-order images, or on a subset of it (region of interest). In one embodiment, the first-order image region of interest is the boundary box of the authentication mark. The boundary box is the smallest convex shape that contains the authentication mark. In another embodiment, an additional set of correlation values is computed based on the so-called DIBS images. The DIBS technique and the meaning of the DIBS images will be apparent from the explanations provided below with reference to FIGS. 24a to 27.

(98) In order to carry out the decision part of step s460 in this implementation, a decision algorithm is used to classify a measured sample into at least two categories: genuine or fake. Known machine learning algorithms may be used for that purpose, such as: support vector machine (SVM), decision trees, K-nearest neighbors algorithm (KNN), etc. In one embodiment, the learning features are the above-described similarity matching values. In one embodiment, other learning features are used, which are not related to cross-correlations such as for example the standard deviation of the pixel values (i.e., intensity values) of the first-order image, or the standard deviation of the pixel values of the zero-order image.

(99) In one embodiment, the standard deviation values and several sets of similarity matching values from images obtained under different excitation wavelengths (e.g. red, green or blue LED) are used. For example, one set of learning features used to describe one sample may be as shown in the following table:

(100) TABLE-US-00001 Feature 1 Value of the correlation of the acquired first-order image with the synthetic first-order image when illuminated by a blue LED Feature 2 Value of the correlation of the first derivative of the acquired first-order image with the first derivative of the synthetic first-order image when illuminated by a blue LED Feature 3 Value of the correlation of the second derivative of the acquired first-order image with the second derivative of the synthetic first-order image when illuminated by a blue LED Feature 4 Value of the correlation of the acquired first-order image with the synthetic first-order image when illuminated by a green LED Feature 5 Value of the correlation of the first derivative of the acquired first-order image with the first derivative of the synthetic first-order image when illuminated by a green LED Feature 6 Value of the correlation of the second derivative of the acquired first-order image with the second derivative of the synthetic first-order image when illuminated by a green LED Feature 7 Value of the standard deviation of the acquired first-order image when illuminated by a blue LED Feature 8 Value of the standard deviation of the acquired first-order image when illuminated by a green LED

(101) In one embodiment, the classifier may be trained in advance on a heterogeneous dataset consisting of randomized genuine samples and non-genuine samples.

(102) During the decision phase of the method, the classifier may classify the given samples using the features input into the classifier.

(103) The above-referred possible implementation of step s400 in the third sub-embodiment has been tested using classification algorithms (in that respect see for example: David Barber, Bayesian Reasoning and Machine Learning, Cambridge University Press 2011) as described in the following table:

(104) TABLE-US-00002 Learning The used features were the correlation value, first features derivative correlation value, second derivative correlation value, all three for blue and green LED excitation, correlation of DIBS values (as discussed below), and first- order standard deviation. Training A KNN classifier was trained on 340 samples (130 genuine and 210 non-genuine). Results When tested against a separate test set of 175 images consisting of never-seen-before backgrounds and codes, the classification accuracy was: 94.29% with 10 false positive and 0 false negative.

(105) Compared with imaging spectrometers used for scientific observations, the approach in the embodiments described with reference to FIGS. 12a to 12c is not focused on reconstructing a hypercube that contains spectral information for every pixel in the acquired image. The approach aims at creating one synthetic image with the assumption that there is only one dominant spectrum involved (the genuine mark spectrum). The computation required to produce that synthetic image consists mainly in several one-dimensional convolutions. In comparison with the computation and memory required to compute a hypercube, the approach is less expensive. Furthermore, the application of a machine learning classifier is also fast and lightweight.

(106) In one embodiment, the convolution or deconvolution operation(s) of step s400 are performed per line of the image along the diffraction direction. Furthermore, when deconvolution step s410 of the embodiment described with reference to FIG. 12a is carried out on a line-by-line manner, the result of the deconvolution may be averaged to reduce noise and cancel possible modulation by the background non-uniformities, prior to comparing the result against the reference spectral information as part of step s420.

(107) In one embodiment, as illustrated by the flowchart of FIG. 13, generating s400 the measure of authenticity further comprises decoding s492 a code from a marking within the imaged non-dispersed part and verifying s494 the authenticity of the code. This enables the snapshot imaging of the marking for decoding s492 and verification s494 (based on the imaged non-dispersed part, i.e. based on the direct image), and then using the output of the code verification process in addition to the verification based on the relation between the imaged dispersed part, the imaged non-dispersed part, and the reference spectral information, to generate the measure of authenticity. For example, in one embodiment, object 10 is regarded as authentic only if both verifications, i.e. the code-based verification and the spectrum-based verification, are successful. In other words, spatial information of a marking or printed code as well as spectral emission information of the marking or printed codewhich may have been printed for example using a photoluminescent inkcan both be obtained for the purpose of authentication.

(108) In one embodiment, the step of decoding s492 the code is used to obtain information based on which the expected spectral composition of the electromagnetic radiation from object 10 and therefore the reference spectral information to be used for the spectrum-based authentication verification in step s400 can be retrieved (e.g. through a database). In such a manner, several different code families each associated with a different ink (and hence a different reference spectrum) can be printed on different classes of products and authenticated with the same device.

(109) In one embodiment, the marking comprises at least one machine readable code, which may for example comprise at least one of a linear barcode and a matrix barcode (e.g., a printed Data Matrix code or QR code). It is therefore possible, in some embodiments of the invention, not only to decode a two-dimensional matrix barcode (or the like) but also to carry out material-based authentication using the spectrum of the radiation coming from object 10, the radiation spectrum corresponding for example to the fluorescence emission spectrum of the ink used for the marking.

(110) In one embodiment, the marking comprises single spectral characteristics at least over one region of the marking. The marking may also comprise single spectral characteristics over the whole marking.

(111) In one embodiment, a mask is intentionally provided, as part of imaging system 200 and in addition thereto, on object 10 or in the vicinity thereof to reveal only a portion of object 10. This is advantageous in the case the whole object carries a substance having the reference spectral information or a large marking which covers the whole image. The mask artificially creates a transition from non-marked to marked area even if there would be no such transition without the mask. In one embodiment, imaging system 200 does not use any slit between dispersive imaging arrangement 30 and object 10. Not using a slit is advantageous in that this enables the simultaneous acquisition of an image and the spectrum thereof, without notably having to scan (by moving the imaging device or spectrometer) the surface of the object to measure the spectrum for each position.

(112) Now, before describing further embodiments of the invention, it may be useful to discuss some of the advantages brought about by some embodiments thereof, especially compared to prior art systems.

(113) The above-described imaging systems and methods in accordance with some embodiments of the invention are advantageous because they allow the construction of simple, compact, snapshot-based (non-scanning), low-cost, and versatile devices, which may for example be incorporated in hand-held audit devices. Acquiring images of both the non-dispersed part of the electromagnetic radiation and the dispersed part thereof indeed suffices, together with the reference spectral information, which is known in advance, to generate the measure of authenticity.

(114) In contrast, imaging spectrometers used for scientific observations, as mentioned above, are typically complex, expensive or bulky. This is because these prior art systems usually aim at obtaining high-resolution spatial and spectral information about all regions of the object or scene.

(115) Mechanical scanning of different bandpass filters in front of an imager allows reconstruction of a spectral irradiance map of the object I(x, y, ). However, the time to scan all filters and the complexity and fragility of the scanning mechanism makes the optical system cumbersome, not rugged and costly to implement.

(116) Tuning systems based on Fabry-Perot interferometer or multistage liquid crystals avoid mechanical complexity but require high-quality and costly optical components (i.e. interferometric mirrors). The scanning of the filter parameters needed to acquire full set of images can be slow and can become another limitation for the use in handheld authentication systems.

(117) Snapshot solutions relying on simultaneous imaging of an object through array of bandpass filters can achieve fast data acquisition and are especially adapted to handheld audit devices. Furthermore, such systems are compact and fit easily in a small volume of a hand-held device. The limited number of different passband filters is, however, a drawback, and it is also difficult to obtain suitable lenslet arrays. In addition, the spectral bands of the filter array have to be optimized to the ink spectral response, which prevents the use of off-the-shelf filter arrays while custom filter arrays are typically expensive to design and manufacture.

(118) The example of a grating-based imager using computer tomography (i.e. CTIS) requires either a complex holographically recorded Kinoform type grating or several crossed gratings able to disperse the light in set of orders around the zero order. The need of several gratings complicates the setup and furthermore, the exposure time should be extended to compensate low efficiency in higher diffraction orders. The data acquisition is therefore slowed, rendering the setup unsuitable for a hand-held device. Such arrangements also require expensive large sensors with multi mega-pixels and extensive calculation for the tomography inversion.

(119) The coded aperture imagers are as slow as the CTIS devices. Moreover, there is an intrinsic problem to reconstruct the full spectrum for specific design of the coded aperture. Meanwhile, integral field spectrometers require cumbersome image slicing optics and require relatively large surface image sensors.

(120) Imaging Fourier transform spectrometers are complex instruments relying on expensive interferometers or birefringent prisms. In either case, the spectrometers require scanning of either an air gap or an angular orientation of the elements to obtain spectra that makes them slow and fragile.

(121) The above-described prior art setups require complex optics and data treatment algorithms to calculate a full spectral data cube I(x, y, ), which is actually not required for authentication purposes. The inventors have found none of these prior art setups suitable for an economical, compact, robust, and fast auditing device based on a spectral imager.

(122) Let us now describe further embodiments of the invention, which may help understand some aspects and advantages of the invention.

(123) In one embodiment, imaging system 200 has an optical setup with a transmission diffraction grating 31 mounted in front of a lens objective 32 in a dispersive imaging arrangement 30 which is then arranged in front of an image sensor arrangement 60, as schematically illustrated on the left-hand side of both FIGS. 14a and 14b. System 200 uses a lens objective 32 of model EO57907 from Edmund Optics Ltd (based in York, UK) with f/2.5 and f=3.6 mm focal length. The dispersive element in arrangement 30 is a transmission diffraction grating 31 of type GT13-06V from Thorlabs, Inc., as mentioned above, with 600 lines-per-mm and 28.7 blaze angle. Area 12 of object 10 is within the field of view of imaging system 200.

(124) FIG. 14a also shows, on the right-hand side of the drawing, the simulated dispersion of a single dot (of, for example, a two-dimensional matrix barcode) at three discrete wavelengths obtained by means of transmission diffraction grating 31 installed in front of imaging objective 32. The dispersion of the diffraction grating 31 obtained from a Zemax OpticStudio simulation is shown. One can see the direct (Order 0) and dispersed images in first positive (Order 1) and first negative (Order 1) orders of the single dot (with diameter of 0.5 mm) onto the image space for three discrete wavelengths.

(125) More complex marks such as full two-dimensional matrix barcodes typically produce smeared images in the first order of the grating 31 due to the specific, broader emission spectra of the inks, and an associated overlap of the successive spread dots in the direction of diffraction is observed, as illustrated on the right-hand side of FIG. 14b. In particular, FIG. 14b shows the simulated dispersion of a data matrix with the non-dispersed image (Order 0) and two images associated with both dispersed orders, i.e. the first positive order (Order 1) and the first negative order (Order 1), assuming equal efficiency of grating 31 for all three orders. The direct image in the zero order of the grating is not influenced by the grating (except for intensity attenuation) and can be used to decode a printed two-dimensional matrix barcode. The scale shown on FIG. 14b is in intensity in arbitrary units (I, a.u.).

(126) Examples of zero- and first-order real images of a two-dimensional matrix barcode printed with two inks, i.e. ink type 1 and ink type 2, are shown in FIG. 15. Namely, FIG. 15 shows the zero- and first-order real images of two-dimensional matrix barcodes printed on labels, with, on the left-hand side of FIG. 15, ink type 1 excited with blue LED light (peak 450 nm), and, on the right-hand side of FIG. 15, ink type 2 excited with red light (peak 640 nm).

(127) It can be observed that the images in zero- and first-order of the grating can be recorded simultaneously (as illustrated by FIG. 35a), sequentially (as illustrated by FIG. 35b), or partially sequentially (as illustrated by FIGS. 35c and 35d), as they both fit on the array detector. Further, the efficiency of the grating for both orders is similar allowing recording both orders with the same exposure time. The efficiency in nth order of a grating is the ratio of diffracted power in the nth order to the total incident power.

(128) The dispersed image in the first order is a convolution (or a convolution-like function) of the zero-order image of the two-dimensional matrix barcode with the ink fluorescence emission spectrum. As a result, the ink emission spectrum can be extracted by deconvolution (or deconvolution-like operation) of the first-order image using the spatial information from the zero-order image that is not affected by the grating dispersion.

(129) A deconvolution algorithm based on fast Fourier transform (FFT) may for example be used to extract the spectrum of the ink. It may use a set of columns from the images, extracted along the grating dispersion direction, comprising intensity profiles from the zero- and first-order images.

(130) FIG. 16 shows an exemplary result (upper-right chart) of the column-by-column deconvolution (or similar non-linear process) from a two-dimensional matrix barcode image containing zero- and first-order components (left-hand image), as well as the comparison of the average of all spectrum curves of the upper-right chart to reference spectral information (lower-right chart), i.e. the spectrum of ink type 1. Each of about 250 columns is subject to deconvolution (or similar non-linear process) and produces a spectrum. The spectra obtained from all columns are then averaged. This averaging reduces the noise (due for example to artefacts that might be created by the deconvolution or similar non-linear process) and cancels local contribution from the background, which may occur on limited parts of the two-dimensional matrix barcode. Hence, the reconstructed spectral profile is an average for the entire printed two-dimensional matrix barcode being observed because it is assumed that all matrix dots are printed with the same ink and there is no significant contribution of background to the emission spectrum.

(131) FIGS. 17 to 19 schematically illustrate three imaging systems 200 in three embodiments of the invention, respectively, showing possible components of dispersive imaging arrangement 30, such as a transmission grating 31, an imaging lens 32, an optical long-pass filter 33, and an additional lens arrangement 34. Area 12 of item 10 can be imaged by arrangement 30, considering its field of view (FOV) 15. Non-dispersed image 41 of area 12 and dispersed image 51 of area 12 corresponding to shortest wavelength are both indicated. Reference 61 is the window 61 of the image sensor(s) 63.

(132) Arrangement 30 of FIG. 17 comprises an imaging lens 32, a transmission grating 31 (600 I/mm) mounted in front of lens 32 (lens objective Edmund Optics 57907), and an optical long-pass filter 33 mounted behind lens 32. As already explained with reference to FIG. 4, this enables to produce low optical aberrations for both dispersed and non-dispersed images by using the broad field-of-view of the lens objective.

(133) Since grating 31 is mounted in front of imaging lens 32, it deflects the beams differently for zero- and first-order and imaging lens 32 receives the input beams at different angles. In such a configuration, a wide-FOV imaging lens 32 is used which allows incident beams at angles specific for the first order.

(134) In arrangement 30 of FIG. 18, both transmission grating 31 (360 I/mm) and optical long-pass filter 33 are mounted behind lens 32 (lens objective Edmund Optics 57907). As already explained with reference to FIG. 5, this enables to cancel the dependence of the extracted spectra on the object position along the optical axis.

(135) In arrangement 30 of FIG. 19, optical long-pass filter 33 is mounted in front of lens 32, and transmission grating 31 (600 I/mm) is mounted behind lens 32 (lens objective Edmund Optics 57907). Furthermore, an additional lens arrangement 34 is also mounted behind lens 32. As already explained with reference to FIG. 6, this configuration enables to efficiently separate the dispersed and non-dispersed images (being approximately double compared to the embodiment of FIG. 18) and avoid dependence on the object position along the optical axis.

(136) Let us now describe further embodiments of the invention involving imaging over a plurality of illumination periods, first with reference to FIGS. 20 and 22 and then with reference to FIGS. 21 and 23a-b. These further embodiments may naturally be combined with any of the above-described embodiments.

(137) FIG. 20 schematically illustrates the generation of a measure of authenticity of object 10 in one embodiment of imaging system 200. In this embodiment, as a first step, image sensor arrangement 60 images the above-described non-dispersed part 40 and dispersed part 50 in a plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n. In one embodiment, n equals 2. In another embodiment, n equals 3. Object 10 is illuminated differently during each illumination period. Each illumination period may encompass one or two imaging periods (either overlapping or non-overlapping), as schematically illustrated with reference to FIGS. 35a to 35d.

(138) Then, the measure of authenticity is generated. The generation of the measure of authenticity comprises the following steps.

(139) First, for each illumination period t.sub.i (1in), an intermediate measure of authenticity k.sub.i is generated depending at least on a relation between dispersed part 50 (A.sub.i) imaged at the illumination period t.sub.i, non-dispersed part 40 (B.sub.i) imaged at the illumination period t.sub.i, and a part of the reference spectral information, said part of the reference spectral information being associated with how object 10 has been illuminated during illumination period t.sub.i. In one embodiment, intermediate measure of authenticity k.sub.i is generated, for each illumination period t.sub.i, by determining, for each illumination period t.sub.i, the extent to which the dispersed part imaged at illumination period t.sub.i corresponds to a convolution of the non-dispersed part imaged at illumination period t.sub.i and said part of the reference spectral information associated with how object 10 has been illuminated during illumination period t.sub.i.

(140) Secondly, the measure of authenticity m is generated based on the plurality of intermediate measures of authenticity k.sub.1, k.sub.2, . . . , k.sub.n. This is illustrated on FIG. 20 by the exemplary equation: m=f(k.sub.1, k.sub.2, . . . , k.sub.n), wherein f is a function such as for example the arithmetic mean of the intermediate measures of authenticity.

(141) FIG. 22 is a flowchart of an imaging method corresponding to the process illustrated by FIG. 20, wherein the generation s400 of the measure of authenticity of object 10 follows image sensor arrangement 60 imaging s300 non-dispersed part 40 and dispersed part 50 in a plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n. The generation s400 of the measure of authenticity comprises generating s470, for each illumination period t.sub.i, an intermediate measure of authenticity k.sub.i as described above, and then generating s475 the measure of authenticity m based on the plurality of generated intermediate measures of authenticity k.sub.1, k.sub.2, . . . , k.sub.n.

(142) In one embodiment, generating s470, for each illumination period t.sub.i, the intermediate measure of authenticity k.sub.i comprises: determining, for each illumination period t.sub.i, the extent to which the dispersed part imaged at illumination period t.sub.i corresponds to a convolution of the non-dispersed part imaged at illumination period t.sub.i and said part of the reference spectral information associated with how object 10 has been illuminated during illumination period t.sub.i.

(143) In one embodiment (not illustrated in FIG. 22), the intermediate measure k.sub.i of authenticity of each illumination period is generated s470 without waiting for the completion of imaging step s300 for all illumination periods. That is, step s470 can be carried out while step s300 is still under way. For example, as soon as image sensor arrangement 60 has imaged non-dispersed part 40 and dispersed part 50 for illumination period t.sub.1, intermediate measure of authenticity k.sub.1 may be generated s470 for illumination period t.sub.1 and then stored, so that generating step s475 may later be carried out based on all stored intermediate measures of authenticity k.sub.1, . . . , k.sub.n.

(144) FIG. 21 schematically illustrates the generation of a measure of authenticity of object 10, in another embodiment of the invention. In this embodiment, as in the embodiment described with reference to FIGS. 20 and 22, image sensor arrangement 60 first images non-dispersed part 40 and dispersed part 50 in a plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n. The value n may for example be equal to 2 or 3, and object 10 is illuminated differently during each illumination period. Again, each illumination period may encompass one or two imaging periods (either overlapping or non-overlapping), as schematically illustrated with reference to FIGS. 35a to 35d. The measure of authenticity is then generated through the following steps:

(145) The imaged non-dispersed part {B.sub.1, B.sub.2, . . . , B.sub.n} is processed based at least on the non-dispersed part B.sub.1 imaged at a first illumination period t.sub.1 among the plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n and the non-dispersed part B.sub.2 imaged at a second illumination period t.sub.2, to produce the processed imaged non-dispersed part B.sub.x. All images B.sub.1, B.sub.2, . . . , B.sub.n may also be taken into account to produce the so-called processed imaged non-dispersed part B.sub.x. That is, the processed imaged non-dispersed part B.sub.x may be generated based on the non-dispersed parts imaged at a first to nth illumination periods t.sub.1, t.sub.2, . . . , t.sub.n. Likewise, the processed imaged dispersed part is generated based at least on the dispersed part A.sub.l imaged at a first illumination period t.sub.1 among the plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n and the dispersed part A.sub.2 imaged at a second illumination period t.sub.2, to produce the so-called processed imaged dispersed part A.sub.x. All dispersed parts A.sub.1, .sub.2, . . . , .sub.n imaged at all the illumination periods t.sub.1, t.sub.2, . . . , t.sub.n may alternatively be taken into account to produce the processed imaged dispersed part A.sub.x.

(146) Then, the measure of authenticity m is generated depending at least on a relation between the processed imaged dispersed part A.sub.x, the processed imaged non-dispersed part B.sub.x, and reference spectral information. In one embodiment, the measure of authenticity m is generated based at least on the extent to which the processed imaged dispersed part A.sub.x corresponds to a convolution of the processed imaged non-dispersed part B.sub.x and reference spectral information.

(147) FIGS. 23a and 23b are two flowcharts of imaging methods in two embodiments corresponding to the process illustrated by FIG. 21, wherein the generation s400 of the measure of authenticity follows image sensor arrangement 60 imaging s300 non-dispersed part 40 and dispersed part 50 in a plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n.

(148) Namely, referring to FIG. 23a, after imaging s300, by image sensor arrangement 60, non-dispersed part 40 and dispersed part 50, in a plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n, the measure of authenticity is generated s400. Step s400 comprises, first, generating s482 the so-called processed imaged non-dispersed part B.sub.x based at least on the non-dispersed parts B.sub.1, B.sub.2 imaged at a first and second illumination period t.sub.1, t.sub.2, and preferably based on all non-dispersed parts B.sub.1, B.sub.2, . . . , B.sub.n imaged at illumination periods t.sub.1, t.sub.2, . . . , t.sub.n. Likewise, the so-called processed imaged dispersed part A.sub.x is generated s484 based at least on the dispersed parts A.sub.1, A.sub.2 imaged at illumination periods t.sub.1, t.sub.2, and preferably based on all non-dispersed parts A.sub.1, . . . , .sub.n imaged at illumination periods t.sub.1, . . . , t.sub.n. Then, the measure of authenticity m is generated s486 depending at least on a relation between processed imaged dispersed part A.sub.x, processed imaged non-dispersed part B.sub.x, and reference spectral information.

(149) In FIG. 23a, steps s482 and s484 are carried out sequentially. However, step s482 may also be carried out after step s484. In one embodiment, steps s482 and s484 are instead carried out in parallel, as illustrated in FIG. 23b.

(150) In one embodiment, step s482 may be implemented as follows (likewise, step s484 may be implemented in a similar manner): First, a weighting factor is calculated based on a statistical processing of pixel values of the first image data B.sub.1 (i.e., the non-dispersed part imaged at illumination period t.sub.1) and pixel values of the second image data B.sub.2 (i.e., the non-dispersed part imaged at illumination period t.sub.2). Then, third image data B.sub.x (i.e., the so-called processed imaged non-dispersed part) is generated by calculating a weighted combination using the pixel values of said first image data B.sub.1, the pixel values of said second image data B.sub.2, and said weighting factor. Such an implementation may be performed to maximize the image contrast between a marking (e.g. a barcode) and the remaining background, as described in PCT application WO 2014/187474 A1 by the same applicant. WO 2014/187474 A1 discloses techniques to enhance the image of a mark or code printed over fluorescing background or other backgrounds. Several images of a mark or code are acquired under different illumination conditions, and an image subtraction algorithm suppresses the background to facilitate the extraction of the printed codes from the images.

(151) This embodiment, which will be described in more detail with reference to FIGS. 24a to 27, can be regarded as a method to enhance the spectral recognition and authentication of a mark (such as for example a printed mark) on backgrounds (such as for example complex fluorescing backgrounds), by using a spectral imager with a dispersive imaging arrangement 30 (such as for example a transmission diffraction grating) and background subtraction using differential images (as described in WO 2014/187474 A1). The background subtraction using differential images, as described in WO 2014/187474 A1, will be hereinafter referred to as the differential illumination background subtraction (DIBS) feature, technique, or algorithm.

(152) This embodiment addresses in particular the following potential problems: The imaged non-dispersed part and imaged dispersed part created by means of dispersive imaging arrangement 30, as discussed above, may overlap and, for example, the fluorescing background of a can cap (or the like) could pose problems for decoding and spectrum extraction. One embodiment of the invention to reduce the effect of overlap is to use optionally an appropriate mask which hides part of the image of object 10 to avoid the overlap between the zero- and first-order images of the code created by means of arrangement 30. Such a mask however is physical and may, under certain circumstances, disturb the code reading by reduction of the useful field of view. Further, the mask may complicate the opto-mechanical design of imaging system 200.

(153) The DIBS-based embodiment aims at addressing such problems. It uses images obtained through arrangement 30 which have an overlap between the orders, and a background subtraction using the DIBS technique is applied. The DIBS technique reduces the effect of fluorescing background (or the like) on the zero-order images (non-dispersed part 40) and further corrects the first-order images (dispersed part 50), thus improving the spectrum-based generation of the measure of authenticity. This is particularly advantageous when the fluorescing background has an excitation spectrum which differs from the ink to be authenticated (e.g. matrix code).

(154) An example of images of a sample object 10 with fluorescing background obtained with an imaging system 200 of FIG. 1 is shown in FIG. 24a (an image of soft-drink can cap without using a mask). A region with overlapping zero- and first-order images of sample object 10 can be observed in FIG. 24a. In this region, it may be difficult or impossible to decode a data matrix due to reduced contrast. This causes the extraction of the spectrum (for generating the measure of authenticity) to be difficult or this may lead to significant errors.

(155) Therefore, the image of FIG. 24a has two problems: 1) the background visible in the zero-order overlaps the first order image, and 2) the background emits light which is diffracted in the 1.sup.st order and interfere spectrally with the spectral information to be authenticated. The first problem may be addressed by using a physical mask. The DIBS technique specifically addresses the second problem, by significantly reducing the background signal from the image.

(156) FIG. 24b shows an image of the same sample object 10 taken with a physical mask in one embodiment of the invention. No overlap between the orders is present which renders an efficient decoding and spectrum extraction possible, but the useful field-of-view may be limited. Such a limitation may, under certain circumstances, restrict the user to operate the device only with specific orientations possibly leading to an increase of authentication time for a sample object 10.

(157) In accordance with the above-mentioned DIBS-based embodiment, no mask is used, but images are acquired in a plurality of illumination periods t.sub.1, t.sub.2, . . . , t.sub.n with several different illuminations and then an image subtraction is carried out in accordance with the DIBS technique. This reduces the influence of a fluorescing background (or the like) on both the decoding (if used) and spectrum extraction.

(158) For example, the DIBS algorithm may use two images acquired by illuminating object 10 with blue and green light respectively. As an output of the algorithm, an image is obtained which is the difference of images taken with blue and green illumination. This image typically has better contrast when it comes to the printed code compared to the initial images, thus improving the performance of the decoding engine (if used). Furthermore, the resulting image also improves the spectrum extraction using the first-order image (i.e., dispersed part 50) created by means of dispersive imaging arrangement 30. This effect may be explained by the different excitation spectra for both the ink used to print the code and the fluorescing background of object 10 (e.g. a soft-drink can cap). The ink is better excited in blue than in green while the background of the soft-drink can cap has mostly the same excitation for both colours. Subtracting the images then leads to increase of the code contrast and improved spectrum extraction.

(159) FIG. 25 shows examples of images of a soft-drink can cap acquired without a physical mask but excited in two different illumination periods by blue light (right-hand image) and green light (left-hand image), in one embodiment of the invention.

(160) FIG. 26 shows examples of background subtracted images using DIBS algorithm, using respectively the linear combinations B0.94*G (right-hand image) and 8.22*(B0.94*G) (left-hand image), in one embodiment of the invention. In the linear combination B0.94*G, B is a first image excited in a first illumination period by blue light, G is a second image excited in a second illumination period by green light, and 0.94 is the weighting factor. In the linear combination 8.22*(B0.94*G), the significance of B, G and 0.94 are the same as for the first linear combination, and 8.22 is a scaling factor. Regarding these linear combinations, the weighting factor and the scaling factor, see equation (1) in WO 2014/187474 A1, page 8, and the corresponding description.

(161) Thanks to the DIBS algorithm, the treated image is more suitable for decoding and improves the spectrum-based generation of the measure of authenticity. FIG. 27 shows examples of extracted spectra with and without DIBS algorithm applied on acquired images in one embodiment of the invention. The extracted spectra can be compared in FIG. 27, where the DIBS pre-treated images allow more precise spectra reconstruction.

(162) Let us now describe further embodiments of the invention applicable to both the imaging over a single illumination period and the imaging over a plurality of illumination periods. These further embodiments may be combined with any of the above-described embodiments.

(163) In one embodiment, object 10 bears a visible or invisible mark (or sign) printed with a printing ink. Such ink contains coloring and/or luminescing agents, such as dye(s) and/or pigment(s) that are typically hard to produce and to reverse-engineer. These optical agents may be classified into two main classes: 1) optical agents producing specific reflective properties upon controlled illumination, and 2) optical agents producing luminescence upon controlled illumination.

(164) The expected spectral response of said optical agents, when subject to particular illumination conditions, is known a priori and constitutes the reference spectral information.

(165) In the case of reflective properties, the spectral response is called the spectral reflectivity, which is the fraction of electromagnetic power reflected per unit of wavelength. For example, FIG. 28 shows the spectral reflectivity of two different color pigments (Microlith from BASF AG, based in Ludwigshafen, Germany), as measured with a spectrophotometer in reflectance mode (e.g. model DU-640 Spectrophotometer from Beckman Coulter Inc., based in Brea, Calif., U.S.).

(166) In order for the reflectivity to be determined, a known broadband illumination source may be used, since the wavelength-dependent reflected electromagnetic radiation 20 (spectral radiance, which is measured) depends on the incident spectral composition of the illumination (spectral irradiance). The spectral reflectivity may be determined either using a calibrated illumination source (in wavelength) or by comparison with a surface of known spectral reflectivity (such as a reference white surface like Spectralon from LabSphere, based in North Sutton, N.H., U.S.) using a non-calibrated broadband light source. The term broadband means that the light source emits at least at all wavelengths in the range of interest. Examples of broadband light source spectral distribution are shown for a white LED (e.g., an OSRAM OSLON SSL white LED) in FIG. 29 and tungsten filament lamp (incandescent bulb) in FIG. 30 (Source: Schroeder, D. V., 2003. Radiant Energy, online chapter for the course, Energy, Entropy, and Everything, Physics Department, Weber State University [accessed May 2016] http://physics.weber.edu/schroeder/eee/chapter6.pdf.).

(167) It can be observed from FIGS. 29 and 30 that the spectrum reflected from a given mark strongly depends on the spectrum of the irradiation source. Therefore, the so-called reference spectral information should be the spectral reflectivity (reflectance) of the object or mark. In embodiments where the reference spectral information is the recorded spectral irradiance, said reference spectral information is then intrinsically related to the spectral distribution of the irradiation source, which should preferably be controlled when the reference spectral information is recorded the first time (enrolled) and also when it is measured to determine the authenticity of object 10.

(168) A second class of optical agents covers luminescent dyes or pigments and has different requirements in terms of illumination and measurement.

(169) Fluorescent dyes and pigments may be selected for example from perylenes (e.g. Lumogen F Yellow 083, Lumogen F Orange 240, Lumogen F Red 300, all available from BASF AG). FIG. 31 (source: WO 2016/042025 A1) shows an example of excitation and emission spectrum of such a fluorescent dye. In particular, it shows the excitation spectrum 601 and emission spectrum 602 of a fluorescent dye (Lumogen F Orange 240 from BASF AG) added in an ink used for printing for example a digital code. Double-headed arrow 603 indicates the wavelength range where the emission spectrum can be used as reference spectral information. It can be observed from FIG. 31 that the excitation spectrum spans between about 400 and 550 nm and the emission spectrum from about 550 to 700 nm. This requires that the illumination source emits at least in the region of excitation for the fluorescent dye to be excited, but preferably not in the emission spectral region to avoid interfering with the fluorescence emission to be detected, which is typically several orders of magnitude weaker than the direct reflection.

(170) This illumination and detection scheme is known in the field of measuring fluorescence and usually comprises a narrow band illumination source such as for example a single color LED (a blue one at 450 nm or a green one at 530 nm may be adapted to excite the Lumogen of FIG. 31) and a long pass optical filter in the detection optical path to cut out any reflection for the tail of the illumination source in the region of emission. Optionally, a short pass optical filter may also be arranged between the LED and the object 10 to be authenticated.

(171) FIGS. 32 and 33 show emission and excitation spectra for two exemplary phosphorescent phosphor pigments: Lumilux blue SN and Lumilux green SN-F2Y from Honeywell International, Inc., based in Morris Plains, N.J., U.S. The spectroscopic properties shown in FIGS. 32 and 33 were measured on samples printed with silk-screen inks using a spectrofluorometer (Horiba Jobin Yvon Fluorolog model FLIII-22, from Horiba, based in Kyoto, Japan). The approach is the same as for the above-described fluorescent dyes or pigments. Excitation spectra 501 and 511 and emission spectra 502 and 522 of two phosphorescent pigments are used for printing marks to be authenticated in the form of patch, logo or designs. Black arrow 505 on each of FIGS. 32 and 33 indicates the wavelength peak of a deep blue LED at 410 nm which may be used for exciting the phosphorescent pigments efficiently.

(172) In one embodiment, the reference spectral information is generated prior to operating the system and method of authentication. This may be done through a recording and registering of the extracted spectral information, in the same or very similar conditions of illumination and detection (for example using the same device or instrument) as the one to be used in the field.

(173) In one embodiment, a non-controlled illumination source may be used, provided that its spectral characteristics can be determined, through a spectral measurement and a subsequent correction may be made prior to extracting the measured spectral information from object 10 or mark to be authenticated.

(174) FIG. 34 is a schematic diagram of an exemplary implementation of a computing unit 700 that may be used in embodiments of the invention, such as, but not only, for generating the above-discussed measure of authenticity.

(175) As illustrated by FIG. 34, a computing unit 700 may include a bus 705, a processing unit 703, a main memory 707, a ROM 708, a storage device 709, an input device 702, an output device 704, and a communication interface 706. Bus 705 may include a path that permits communication among the components of computing unit 700.

(176) Processing unit 703 may include a processor, a microprocessor, or processing logic that may interpret and execute instructions. Main memory 707 may include a RAM or another type of dynamic storage device that may store information and instructions for execution by processing unit 703. ROM 708 may include a ROM device or another type of static storage device that may store static information and instructions for use by processing unit 703. Storage device 709 may include a magnetic and/or optical recording medium and its corresponding drive.

(177) Input device 702 may include a mechanism that permits an operator to input information to processing unit 703, such as a keypad, a keyboard, a mouse, a pen, voice recognition and/or biometric mechanisms, etc. Output device 704 may include a mechanism that outputs information to the operator, including a display, a printer, a speaker, etc. Communication interface 706 may include any transceiver-like mechanism that enables computing unit 700 to communicate with other devices and/or systems (such as with a base station, a WLAN access point, etc.). For example, communication interface 706 may include mechanisms for communicating with another device or system via a network.

(178) Computing unit 700 may perform certain operations or processes described herein. These operations may be performed in response to processing unit 703 executing software instructions contained in a computer-readable medium, such as main memory 707, ROM 708, and/or storage device 709. A computer-readable medium may be defined as a physical or a logical memory device. For example, a logical memory device may include memory space within a single physical memory device or distributed across multiple physical memory devices. Each of main memory 707, ROM 708 and storage device 709 may include computer-readable media. The magnetic and/or optical recording media (e.g., readable CDs or DVDs) of storage device 709 may also include computer-readable media. The software instructions may be read into main memory 707 from another computer-readable medium, such as storage device 709, or from another device via communication interface 706.

(179) The software instructions contained in main memory 709 may cause processing unit 703 to perform operations or processes described herein, such as for example generating the measure of authenticity. Alternatively, hardwired circuitry may be used in place of or in combination with software instructions to implement processes and/or operations described herein. Thus, implementations described herein are not limited to any specific combination of hardware and software.

(180) FIGS. 35a to 35d schematically illustrate examples of imaging period(s) and illumination period, in four embodiments of the invention. These drawings have been already referred to and elaborated upon throughout the above description.

(181) In one embodiment, imaging system 200 comprises, on the one hand, an imaging device comprising image sensor arrangement 60 and, on the other hand, a piece of equipment, hereinafter referred to as imaging accessory, comprising dispersive imaging arrangement 30.

(182) In this embodiment, the imaging device has a built-in camera (including associated lenses) and may be a hand-held device, such as for example at least one of: a mobile phone, a smartphone, a feature phone, a tablet computer, a phablet, a portable media player, a netbook, a gaming device, a personal digital assistant, and a portable computer device. The imaging device's built-in camera image sensors act as image sensor arrangement 60 in system 200.

(183) As mentioned above, the imaging accessory comprises dispersive imaging arrangement 30, such as for example a transmission diffraction grating, or any other dispersive element as already discussed above with reference to FIG. 1.

(184) The imaging accessory is attachable, directly or indirectly (for example via a connecting piece of equipment), to the imaging device so that the imaging accessory's dispersive imaging arrangement 30 is positioned relative to the imaging device's image sensor arrangement 60 in such a manner that the imaging device and the imaging accessory form an imaging system 200 as described above, operable for imaging an object and generating a measure of authenticity of the object. In other words, the imaging accessory may be used for example to transform a smartphone into a portable imaging and authentication system as described above. The imaging accessory may for example be fixedly positionable over the smartphone rear camera. The processing and communications capabilities of the smartphone may then be used for implementing a processing unit 70 of imaging system 200.

(185) Furthermore, if the imaging device has a light source (such as for example flash LEDs used in a smartphone), said light source may operate as illumination arrangement 210 to illuminate the object 10 to be imaged and authenticated. A smartphone's light source is typically well adapted for reflectivity measurements. Alternatively, illumination arrangement 210 may be provided as part of the imaging accessory.

(186) This embodiment is advantageous in that the imaging accessory may be a passive accessory, requiring no additional power, and thus providing an affordable authentication solution.

(187) FIG. 36 schematically illustrates an imaging system 200 in accordance with the above-described embodiment comprising, on the one hand, an imaging device comprising image sensor arrangement 60, wherein the imaging device is a mobile phone having a camera, and, on the other hand, an imaging accessory 36 comprising dispersive imaging arrangement 30. In this exemplary optical setup, imaging accessory 36 comprises a diffraction grating 31 and long pass filter 33 arranged in front of the mobile phone camera 64. The mobile phone camera 64 comprises an image sensor 60 and a built-in lens 66. Optionally, an additional collimating lens 35 may be positioned in front of imaging accessory 36.

(188) The invention further relates to the following embodiments: Embodiment (X2). Imaging system (200) of claim 1, wherein the imaging system (200) is an imaging device. Embodiment (X3). Imaging system (200) of claim 1, comprising an imaging device (100) comprising the image sensor arrangement (60) and the dispersive imaging arrangement (30), wherein the imaging device (100) is not configured to generate the measure of authenticity. Embodiment (X4). Imaging system (200) of embodiment (X2) or (X3), wherein the imaging device is a hand-held device. Embodiment (X7). Imaging system (200) according to any one of claims 1 to 3 and embodiments (X2) to (X4), wherein the imaging system (200) is configured for generating the measure of authenticity after the image sensor arrangement (60) has, in a plurality of illumination periods (t.sub.1, t.sub.2, . . . , t.sub.n), imaged the non-dispersed part (40) and the dispersed part (50); and generating the measure of authenticity comprises: generating, for each illumination period (t.sub.i), an intermediate measure of authenticity (k.sub.i) depending at least on a relation between the dispersed part imaged at the illumination period (t.sub.i), the non-dispersed part imaged at the illumination period (t.sub.i), and a part of the reference spectral information, said part of the reference spectral information being associated with how the object (10) has been illuminated during the illumination period (t.sub.i); and generating the measure of authenticity (m) based on the plurality of generated intermediate measures of authenticity (k.sub.1, k.sub.2, . . . , k.sub.n). Embodiment (X8). Imaging system (200) of embodiment (X7), wherein generating, for each illumination period (t.sub.i), the intermediate measure of authenticity (k.sub.i) comprises: determining, for each illumination period (t.sub.i), the extent to which the dispersed part imaged at the illumination period (t.sub.i) corresponds to a convolution of the non-dispersed part imaged at the illumination period (t.sub.i) and said part of the reference spectral information associated with how the object (10) has been illuminated during the illumination period (t.sub.i). Embodiment (X9). Imaging system (200) according to any one of claims 1 to 3 and embodiments (X2) to (X4), wherein the imaging system (200) is configured for generating the measure of authenticity after the image sensor arrangement (60) has, in a plurality of illumination periods (t.sub.1, t.sub.2, . . . , t.sub.n), imaged the non-dispersed part (40) and the dispersed part (50); and generating the measure of authenticity comprises: processing the imaged non-dispersed part based at least on the non-dispersed part imaged at a first illumination period (t.sub.1) among the plurality of illumination periods (t.sub.1, t.sub.2, . . . , t.sub.n) and the non-dispersed part imaged at a second illumination period (t.sub.2) among the plurality of illumination periods (t.sub.1, t.sub.2, . . . , t.sub.n), wherein the illumination conditions during the first illumination period (t.sub.1) at least partially differ from the illumination conditions during the second illumination period (t.sub.2); processing the imaged dispersed part based at least on the dispersed part imaged at the first illumination period (t.sub.1) and the dispersed part imaged at the second illumination period (t.sub.2); and generating the measure of authenticity (m) depending at least on a relation between the processed imaged dispersed part (A.sub.x), the processed imaged non-dispersed part (B.sub.x), and the reference spectral information. Embodiment (X10). Imaging system (200) of embodiment (X9), wherein generating the measure of authenticity (m) depends at least on the extent to which the processed imaged dispersed part (A.sub.x) corresponds to a convolution of the processed imaged non-dispersed part (B.sub.x) and the reference spectral information. Embodiment (X11). Imaging system (200) according to any one of claims 1 to 3 and embodiments (X2) to (X4) and (X7) to (X10), wherein the dispersive imaging arrangement (30) is positioned relative to the image sensor arrangement (60) in such a manner as to allow the image sensor arrangement (60) to image the non-dispersed part (40) and the dispersed part (50) in two portions of the same image sensor. Embodiment (X13). Imaging system (200) according to any one of claims 1 to 4 and embodiments (X2) to (X4) and (X7) to (X11), wherein a slit is not used between the dispersive imaging arrangement (30) and the object (10) to be imaged. Embodiment (X17). Imaging system (200) of claim 7, wherein the at least one machine readable code comprises at least one of a linear barcode and a matrix barcode. Embodiment (X18). Imaging system (200) according to any one of claims 5 to 7 and embodiment (X17), wherein the marking (11) comprises single spectral characteristics at least over one region of the marking (11). Embodiment (X19). Imaging system (200) of embodiment (X18), wherein the marking (11) comprises single spectral characteristics over the whole marking (11). Embodiment (X20). Imaging system (200) according to any one of claims 5 to 7 and embodiments (X17) to (X19), wherein the marking (11) comprises at least one of: optical agents producing specific reflective properties upon controlled illumination, and optical agents producing luminescence upon controlled illumination. Embodiment (X21). System (220) comprising an imaging system (200) according to any one of claims 1 to 7 and embodiments (X2) to (X4), (X7) to (X11), (X13), and (X17) to (X20), and an illumination arrangement (210) for controlled illumination of the object (10). Embodiment (X23). Imaging method of claim 8, wherein the imaging method is carried out by an imaging device. Embodiment (X24). Imaging method of claim 8, wherein the imaging method is carried out by an imaging system (200) comprising an imaging device (100) comprising the image sensor arrangement (60) and the dispersive imaging arrangement (30), wherein the imaging device (100) does not generate (s400) the measure of authenticity. Embodiment (X25). Imaging method of embodiments (X23) or (X24), wherein the imaging device is a hand-held device. Embodiment (X32). Imaging method according to any one of claims 8 to 14 and embodiments (X23) to (X25), wherein the dispersive imaging arrangement (30) is positioned relative to the image sensor arrangement (60) in such a manner as to allow the image sensor arrangement (60) to image the non-dispersed part (40) and the dispersed part (50) in two portions of the same image sensor. Embodiment (X33). Imaging method according to any one of claims 8 to 14 and embodiments (X23) to (X25) and (X32), wherein the dispersive imaging arrangement (30) comprises at least one of: a diffractive element, a transmission diffraction grating, a blazed transmission diffraction grating, a volume holographic grating, a reflective diffraction grating,

(189) an arrangement comprising a beam splitter and a diffraction grating, and an arrangement comprising a beam splitter and a dispersive prism. Embodiment (X34). Imaging method according to any one of claims 8 to 14 and embodiments (X23) to (X25), (X32) and (X33), wherein a slit is not used between the dispersive imaging arrangement (30) and the object (10) to be imaged. Embodiment (X37). Imaging method of claim 15 or 16, wherein the marking (11) comprises at least one machine readable code. Embodiment (X38). Imaging method of embodiment (X37), wherein the at least one machine readable code comprises at least one of a linear barcode and a matrix barcode. Embodiment (X39). Imaging method according to any one of claims 15 and 16 and embodiments (X37) and (X38), wherein the marking (11) comprises single spectral characteristics at least over one region of the marking (11). Embodiment (X40). Imaging method of embodiment (X39), wherein the marking (11) comprises single spectral characteristics over the whole marking (11). Embodiment (X41). Imaging method according to any one of claims 15 and 16 and embodiments (X37) and (X40), wherein the marking (11) comprises at least one of: optical agents producing specific reflective properties upon controlled illumination, and optical agents producing luminescence upon controlled illumination. Embodiment (X43). Computer program or set of computer programs comprising computer-executable instructions configured, when executed a computer or set of computers, to carry out an imaging method according to any one of claims 8 to 16 and embodiments (X23) to (X25), (X32) to (X34), and (X37) to (X41). Embodiment (X44). Computer program product or set of computer program products comprising a computer program or set of computer programs according to embodiment (X43). Embodiment (X45). Storage medium storing a computer program or set of computer programs according to embodiment (X43).

(190) Where the terms processing unit, storage unit, etc. are used herewith, no restriction is made regarding how distributed these elements may be and regarding how gathered elements may be. That is, the constituent elements of a unit may be distributed in different software or hardware components or devices for bringing about the intended function. A plurality of distinct elements may also be gathered for providing the intended functionalities.

(191) Any one of the above-referred units, such as for example processing unit 70, or devices, such as for example imaging device 110, may be implemented in hardware, software, field-programmable gate array (FPGA), application-specific integrated circuit (ASICs), firmware or the like.

(192) In further embodiments of the invention, any one of the above-mentioned processing unit, storage unit, etc. is replaced by processing means, storage means, etc. or processing module, storage module, etc. respectively, for performing the functions of the processing unit, storage unit, etc.

(193) In further embodiments of the invention, any one of the above-described procedures, steps or processes may be implemented using computer-executable instructions, for example in the form of computer-executable procedures, methods or the like, in any kind of computer languages, and/or in the form of embedded software on firmware, integrated circuits or the like.

(194) Although the present invention has been described on the basis of detailed examples, the detailed examples only serve to provide the skilled person with a better understanding, and are not intended to limit the scope of the invention. The scope of the invention is much rather defined by the appended claims. Abbreviations: ASICs application-specific integrated circuit a.u. arbitrary units CASSI coded aperture snapshot spectral imager CCD charge-coupled device CMOS complementary metal-oxide-semiconductor CTIS computed tomography imaging spectrometer DIBS differential illumination background subtraction FOV field of view FPGA field-programmable gate array KNN K-nearest neighbors algorithm I/mm lines per mm LED light-emitting diode LTI linear translation-invariant MAFC multi-aperture filtered camera MIFTS multiple-image Fourier transform spectrometer NIR near-infrared RAM random-access memory ROM read-only memory SHIFT snapshot hyperspectral imaging Fourier transform spectrometer SVM support vector machine SWIR short-wavelength infrared UV ultraviolet WLAN wireless local area network