COHERENT LIDAR IMAGING SYSTEM
20230048766 · 2023-02-16
Assignee
Inventors
Cpc classification
G01S7/4917
PHYSICS
G01S17/894
PHYSICS
G01S17/34
PHYSICS
International classification
G01S17/34
PHYSICS
G01S7/481
PHYSICS
Abstract
A coherent LIDAR imaging system includes a laser source; an optical splitter/recombiner designed to split the laser radiation into a reference beam and into an object beam and to superpose the reference beam on a reflected object beam reflected by the scene; and an optical imager creating an image of the scene on a detector. The detector includes an array of pixels designed for detecting the reflected object beam and the reference beam which together form a recombined beam having a beat frequency representative of a range of the illuminated scene. The optical splitter/recombiner is configured to form an intermediate image of the reference beam in an intermediate image plane perpendicular to the optical axis.
Claims
1. A coherent LIDAR imaging system comprising: a laser source configured to emit laser radiation (L) having a time modulated optical frequency; an optical splitter/recombiner comprising at least: an optical splitter configured to spatially split the laser radiation into a reference beam directed toward an optical imager and into an object beam directed toward a scene that is to be illuminated; and an optical recombiner configured to spatially superpose the reference beam on part of the object beam reflected by the scene, referred to as reflected object beam; the optical imager having an optical axis creating an image of the illuminated scene on a detector by focusing the reflected object beam, the detector comprising an array of pixels, each of the pixels comprising a photodetector configured to detect a portion of the reflected object beam and a portion of the reference beam which together form a portion of a recombined beam having a beat frequency representative of a range of the illuminated scene; and a processor connected to the detector and to the laser source and configured to determine a range of points in the scene, imaged on the pixels and to construct a range image of the scene, from the beat frequency associated with each pixel and from the modulated optical frequency of the laser radiation; wherein: the optical splitter/recombiner is further configured to form an intermediate image of the reference beam in an intermediate image plane perpendicular to the optical axis, the position of the intermediate image plane on the optical axis being comprised in a segment along the optical axis corresponding to a projection onto the optical axis of a zone of intersection of subbeams of the reflected object beam, the subbeams being full-light edge-of-field beams.
2. The imaging system as claimed in claim 1, wherein the optical splitter/recombiner is configured so that the intermediate image plane is coincident with a plane containing a pupil or an aperture diaphragm of the optical imager.
3. The imaging system as claimed in claim 1, wherein the optical splitter/recombiner is configured so that a lateral dimension of the intermediate image is less than or equal to a lateral dimension of the intersection of the subbeams in the intermediate image plane.
4. The imaging system as claimed in claim 1, wherein the optical splitter/recombiner is configured so that a divergence of the reference beam coming from the optical splitter/recombiner is equal to an angle formed by the subbeams in the intermediate image plane.
5. The imaging system as claimed in claim 1 wherein the optical splitter/recombiner is configured so that, for each pixel, an axis of propagation of the portion of the reflected object beam forms, with an axis of propagation of the portion of the reference beam, an angle θ such that θ<λ/(2×a.sub.pix), where a.sub.pix is a characteristic dimension of the photodetector of each pixel of the detector and X is a wavelength of the laser radiation.
6. The imaging system as claimed in claim 1, comprising an optical shaper placed in a path of the laser radiation or of the object beam and configured so that a beam coming from the optical shaper displays uniform illumination of the scene that is to be observed.
7. The imaging system as claimed in claim 1, wherein the optical imager has a numerical aperture such that a diameter of speckles on the detector is greater than a characteristic dimension a.sub.pix of the photodetector of each pixel of the detector.
8. The imaging system as claimed in claim 1, wherein the laser source is configured so that the optical frequency is modulated by a periodic linear ramp of excursion B and duration T and so that a coherence length of the laser radiation is twice as high as a predetermined maximum distance z.sub.max between the scene that is to be illuminated and the imaging system.
9. The imaging system as claimed in claim 8, wherein each pixel comprises an electronic processing circuit designed to calculate the beat frequency of the portion of the recombined beam.
10. The imaging system as claimed in claim 9, wherein, with the calculation of the beat frequency being performed simultaneously for all the pixels of the detector, each pixel comprising an interconnection layer in a dielectric material transparent to the recombined beam, the interconnection layer comprising metallic interconnections connected to the electronic processing circuit, the interconnection layer being arranged on a side of the photodetector opposite to a side detecting the portion of the recombined beam, the electronic circuit associated with the pixel being configured to: collect a photocurrent generated in the photodetector representative of the beat frequency and convert the photocurrent into a voltage, then substitute a DC component of the signal representative of the beat frequency with a predetermined fixed voltage value, generating a substituted signal; then amplify the substituted signal, generating an amplified signal; detect, using a comparator, peaks or rising fronts or falling fronts of the amplified signal, and proceed to count a number of detected periods, and transmit to the processor a signal representative of the count of the number of detected periods, the processor determining a range of a point in the scene by calculating a number N of periods detected over the duration T of the optical frequency modulation from the signal representative of the count.
11. The imaging system as claimed in claim 9, wherein a pixel sampling frequency is greater than two times a frequency F.sub.R,max corresponding to the beat frequency associated with an object in the scene which is situated at a maximum predetermined range z.sub.max, and such that F.sub.R,max=2Bz.sub.max/cT, where c is the speed of light.
12. The imaging system as claimed in claim 1, wherein the optical splitter comprises a first integrated optical circuit in which the laser radiation is coupled, at least one waveguide of the first integrated circuit guiding the reference beam, the first integrated optical circuit further comprising a plurality of waveguides each comprising at least one diffraction grating so as to form the object beam.
13. The imaging system as claimed in claim 12, wherein the optical recombiner comprises a second integrated optical circuit, in which the reference beam is coupled by means of an optical element, the second integrated optical circuit comprising at least one waveguide comprising at least one reference grating, the reference grating coupling the reference beam to free space and to the optical imager (Im), the second integrated optical circuit being positioned in the intermediate image plane so that the reference grating forms the intermediate image.
14. The imaging system as claimed in claim 13, wherein the optical element is an optical fiber, guiding the reference beam coming from the first integrated optical circuit toward the second integrated optical circuit.
15. The imaging system as claimed in claim 1, wherein the optical splitter/recombiner comprises an integrated optical circuit in which the laser radiation is coupled, the integrated optical circuit comprising the optical splitter and the optical recombiner, the optical splitter comprising a plurality of waveguides of the integrated optical circuit and each comprising at least one diffraction grating, referred to as object grating, the object gratings coupling the object beam to free space and to the scene that is to be observed, and at least one reference waveguide of the integrated circuit guiding the reference beam toward the optical recombiner comprising at least one reference grating situated on the reference waveguide, the integrated optical circuit being arranged in the intermediate image plane so that the reference grating forms the intermediate image.
16. The imaging system as claimed in claim 1, wherein the optical splitter/recombiner further comprises an intermediate optical device, positioned after the optical splitter, and before the optical recombiner, the intermediate optical device being designed to focus the reference beam and form the intermediate image.
17. An imaging method using a coherent LIDAR imaging system, comprising the following steps: emitting laser radiation from a laser source having a time-modulated optical frequency (F.sub.L); spatially separating, using an optical splitter of an optical splitter/recombiner, the laser radiation into a reference beam and into an object beam directed toward a scene that is to be illuminated; spatially superposing, using an optical recombiner of the optical splitter/recombiner, a reflected object beam reflected by the scene and the reference beam; creating, using an optical imager, an image of the scene on a detector by focusing the reflected object beam, the detector comprising an array of a plurality of pixels; detecting, on each of the pixels, a portion of the reflected object beam and a portion of the reference beam which together form a portion of a recombined beam having a beat frequency representative of a range of the illuminated scene; the optical splitter/recombiner being configured to form an image-intermediate image of the reference beam in an intermediate image plane perpendicular to an optical axis, the position of the intermediate image plane on the optical axis being comprised in a segment along the optical axis corresponding to a projection onto the optical axis of a zone of intersection between subbeams of the reflected object beam, said subbeams being full-light edge-of-field beams; and determining, using a processor connected to the detector and to the laser source, a range to points of the scene which are imaged on the pixels and constructing a range image of the scene from the beat frequency associated with each pixel and from the modulated optical frequency of the laser radiation.
Description
BRIEF DESCRIPTION OF THE FIGURES
[0051] Further features, details and advantages of the invention will become apparent from reading the description given with reference to the attached drawings provided by way of example and which, respectively, depict:
[0052]
[0053]
[0054]
[0055]
[0060]
[0061]
[0062]
[0063]
[0064]
[0065]
[0066]
[0067]
[0068]
DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS
[0069] In the figures and the remainder of the description, the same references represents elements that are identical or similar. In addition, the various elements are not depicted to scale in order to make the figures clearer. Furthermore, the various embodiments and variants are not exclusive of one another and may be combined with one another. Unless indicated otherwise, the terms “substantially”, “approximately” and “of the order of” mean to within 10% and preferably to within 5%. Furthermore, terms “of between . . . and . . . ” and equivalents mean that the end points are included, unless indicated otherwise.
[0070]
[0071] Note that the imaging system D is said to be of the LIDAR (Light Detection and Ranging) type insofar as it is designed to provide a map of the range (or a range image) of a scene illuminated by a coherent optical signal. In the context of the invention, the imaging system illuminates the entire scene instantaneously using a continuous optical signal and simultaneously acquires the image of the illuminated scene in order to determine the range thereof. The imaging system according to the invention thus differs from LIDAR systems (of the time of flight or TOF measurement type or even of the FMCW type) in which the scene is “scanned”, i.e. spatially scanned by an optical signal.
[0072] Note too that the imaging system D is said to be a heterodyne detection system insofar as, in order to determine the range of the illuminated scene, a frequency referred to as the beat signal of a heterodyne signal (recombined beam) formed by the interference between a reference beam of a local oscillator and a beam backscattered/reflected by the illuminated scene is determined, these two optical beams being coherent with one another. The reference beam and the beam projected onto the scene effectively both come from the same primary optical signal (laser radiation L) emitted by the laser source. The heterodyne detection is of FMCW type insofar as the primary signal is a frequency-modulated continuous signal.
[0073] The imager (detector Det) may be of the global shutter type in which all the pixels P.sub.i are read at the same instant. The imaging system is thus able to provide an instantaneous range image. As a variant, it may be of the rolling shutter type and therefore offer sequential reading of the pixels P.sub.i, for example row by row, where all the pixels P.sub.i of the one same row are read simultaneously.
[0074] The imaging system D comprises a laser source SL configured to emit laser radiation L with a time-modulated optical frequency F.sub.L. The laser source is configured so that a coherent length of the source is two times greater, preferably 10 times greater than a predetermined maximum range z.sub.max between the most distant object in the scene and the imaging system D. Specifically, as the imaging system D is a coherent LIDAR system, the coherence length I.sub.C of the laser source is greater than the difference in optical pathway between the object path and the reference path—which is approximately the there-and-back distance between the scene that is to be observed Obj and the imaging system D (minus the optical pathway of the reference beam between the splitter blade and the recombiner blade)—in order to perform coherent detection.
[0075] The reference path is the pathway followed by the reference beam L.sub.ref between the laser source SL and the detector array Det. The object path is the outbound path followed by the object beam L.sub.o from the laser source SL to the scene, plus the return pathway of the reflected/backscattered object beam L.sub.or reflected/backscattered by the scene as far as the detector array Det.
[0076] The laser radiation L emitted by the laser source SL is a coherent, continuous and frequency-modulated primary signal. It is frequency modulated, for example in this instance linearly, on the basis of a starting frequency fo over a repetition period T with a bandwidth B. The signal in this instance is a chirp signal, which is to say a sinusoidal wave of which the instantaneous frequency changes linearly with time.
[0077] The laser source SL may, in the case of emission in the near infrared domain (between 0.7 and 2 μm) be of the Vertical-Cavity Surface Emitting Laser (VCSEL) type which generally has a coherence length of the order of one meter, or may even be of the Edge Emitting Laser (EEL) type which may have a coherence length of the order of a few tens or hundreds of meters.
[0078] Passive optical elements may be situated on the optical pathway of the laser radiation L (primary signal) between the laser source SL and the optical splitter/recombiner SSR. Thus, a shaping lens may be situated at the output of the laser source SL and allows the optical beam to be collimated while broadening its lateral dimension, for example to a diameter of a few millimeters. In addition, a spatial filtering device may be present to eliminate the high spatial frequencies. Thus, the laser radiation L can propagate in a collimated manner, with possibly a gaussian profile and a diameter of the order of a few millimeters, for example 5 mm.
[0079] The imaging system D comprises an optical splitter/recombiner SSR comprising at least: [0080] an optical splitter LS designed to spatially split the laser radiation L into two paths: a reference path with which there is associated a reference beam L.sub.ref and an object path with which there is associated an object beam L.sub.o directed toward the scene that is to be illuminated Obj; [0081] an optical recombiner SR designed to spatially superpose the reference beam L.sub.ref on an object beam reflected by the scene, referred to as reflected object beam L.sub.or so as to form a recombined beam L.sub.rec (heterodyne signal).
[0082] According to one embodiment of the invention, the optical splitter is configured so that the distribution of the intensity of the light radiation of the reference beam L.sub.ref and of the object beam L.sub.o are approximately equal to 10%/90% (at the output of the optical splitter LS). Other values for intensity ratio may be selected in order to optimize the photometric performance of the imaging system D, notably depending on the characteristics of the scene (for example Lambertian or specular reflection) and of the imaging system D (for example the splitting and recombining factors of the optical splitter and recombiner).
[0083] The imaging system D is thus designed to project part of the laser radiation L toward a scene that is to be illuminated instantaneously. What is meant by “scene” is the scene illuminated by the object beam L.sub.o. Thus, the entire scene is illuminated instantaneously by the object beam L.sub.o which therefore has an angular aperture capable of illuminating the entire scene instantaneously. As a preference, the scene is also illuminated uniformly, which is to say that the object beam L.sub.o has an intensity the angular distribution of which is constant.
[0084] Thus, with a view to illuminating the scene instantaneously and, in this instance, uniformly, the imaging system D may perform shaping of the object beam L.sub.o. To do that, it may comprise at least one optical element designed to make the object beam L.sub.o divergent. The angular aperture of the object beam L.sub.o is such that the entire scene is illuminated uniformly by the object beam L.sub.o. It is preferably substantially equal to the field angle of the optical imager Im (for example equal to approximately 20°). In addition, it may comprise at least one optical element designed to render the angular distribution of the intensity constant, in order to obtain a flat intensity profile rather than a gaussian profile. Of course, the various optical elements may be separate or combined.
[0085] The imaging system D further comprises an optical imager Im designed to perform the imaging of the scene that is to be observed Obj on a detector Det by focusing the reflected object beam L.sub.o,r. The detector Det is positioned approximately in the image focal plane of the optical imager Im. In other words, the scene is illuminated and the received plane of the detector array Det are conjugated, give or take the depth of the field. According to one embodiment, the optical imager Im has a focal length that is much shorter than the typical range between the scene and the detector Det so that the scene can be considered to lie at infinity from the view point of the optical imager Im. The optical imager Im is an objective lens comprising an aperture diaphragm Diaph that defines the physical pupil of the system. As detailed later on, it allows a coherent spatial superposition of the reflected object beam L.sub.or with the reference beam L.sub.ref (corresponding to the local oscillator) on the detector Det, thus improving the quality of the detected heterodyne signal. Note that the imaging device Im allows the use of a common path in order to create an image of the scene and cause the object path and the reference path to interfere. This allows the imaging system D to be far more compact in comparison with LIDARs of the prior art that employ waveguide recombination.
[0086] The optical imager Im has optical elements in free space, and comprises at least one lens and an aperture diaphragm which defines the physical pupil. Note that the aperture diaphragm does not have to be a physical object distinct from a lens, but may be defined by the contour of the lens. The optical imager Im comprises an optical axis AO orthogonal to the received plane of the detector array Det. In this instance it is designed to receive the reflected object beam L.sub.or and the reference beam L.sub.ref so that these can be transmitted to the detector array Det along the one same optical axis AO thus allowing better superposition of the two optical beams, thus improving the combination of the two optical beams by interference, making it possible to improve the intensity of the heterodyne signal.
[0087] According to one embodiment, the optical imager Im is formed of a lens or a plurality of lenses, for example a triple Gauss (cf.
[0088] The imaging system D is designed so that all the pixels P.sub.i of the detector array Det receive the reference beam L.sub.ref. This goes hand in glove with the fact that all the pixels P.sub.i also receive the reflected object signal L.sub.or. In this way, the recombined beam L.sub.rec (heterodyne signal) is formed at each pixel P.sub.i so that a range image can be determined from the entire array of the pixels P.sub.i. By way of example, a diffuser and possibly a convergent lens allow the reference beam L.sub.ref to be shaped according to a predefined angular aperture. These optical elements are then positioned between the optical device LS and SR. The angular aperture is predefined to take account of the various optical elements of the optical imager Im, so that the reference beam L.sub.ref effectively illuminates each pixel P.sub.i. Note that the convergent lens between the systems LS and SR, downstream of the diffuser, is not needed: all that is required is for the diffuser to be correctly positioned in the path of the reference beam L.sub.ref, taking account of its diffusion cone, of the diameter of the reference beam L.sub.ref incident on the diffuser, of the aperture of the optical imager Im and of the size of the array of pixels P.sub.i of the detector Det. In a variant (guided optics configuration described later on), diffraction gratings may be dimensioned so that the reference beam L.sub.ref has an angular aperture allowing illumination of all the pixels P.sub.i.
[0089] The diffuser situated between the systems LS and SR and the received plane of the detector array Det are not conjugate, and this then prevents the image of the diffuser from being formed on the detector array Det, as such an image could impair the quality of the heterodyne signal to be detected. By contrast, the diffuser may be conjugate with a plane referred to as intermediate plane situated at or near the aperture diaphragm of the optical imager Im, so as to form a light source equivalent to the reference beam L.sub.ref as explained in detail later on.
[0090] The detector Det is a detector array. It comprises an array of a plurality of pixels P.sub.i, with i ranging from 1 to n. Each of the pixels P.sub.i comprises a photodetector Ph designed to detect a portion L.sub.or,i of the reflected object beam L.sub.or and a portion L.sub.ref,i of the reference beam L.sub.ref, both focused by the optical imager Im on the pixel P.sub.i. The received plane of the detector array Det is situated in a plane conjugate with the scene by the optical imager Im (give or take the depth of field). In other words, the image of the scene is formed in the receiving plane of the detector array Det. Each pixel P.sub.i is intended to receive the reference beam L.sub.ref and the reflected object beam L.sub.or, which interfere with one another to form a signal referred to as a heterodyne signal (recombined beam L.sub.rec). The detector may be of CMOS or CCD type.
[0091] Note that the heterodyne signal comprises a constant component DC.sup.het (which is intended to be filtered out) and a periodic alternating component AC.sup.het(t) which forms the useful signal. This alternating component has a beat frequency here equal to BT/T and therefore equal to 2BD/cT, c being the speed of light. In this instance, τ is a delay between the reflected object beam L.sub.or and the reference beam L.sub.ref, B is the passband and T is the period of the frequency modulation of the primary signal (laser radiation L). Determining the beat frequency therefore makes it possible to obtain the range between the scene and the detector array Det.
[0092] The pixels P.sub.i may be formed of a microlens positioned facing a photodiode. Each pixel P.sub.i has a lateral dimension d.sub.px defined by the size of the microlenses, whereas the photosensitive zone of the photodiodes has a lateral dimension d.sub.pd. Here, it is considered that the detection plane is the plane in which the microlenses extend.
[0093] Each pixel P.sub.i of the detector Det may comprise an electronic processing circuit CE designed to calculate the beat frequency F.sub.i of the heterodyne signal, namely of the recombined beam L.sub.rec formed of the interference between the detected portion of the reflected object beam L.sub.or,i with the detected portion of the reference beam L.sub.ref,i. The calculation of the beat frequency is preferably simultaneous for all the pixels P.sub.i of the detector Det (see later). That makes it possible to create an instantaneous range image of the scene (“global shutter”) without the need to scan the scene point by point.
[0094] For that, each pixel P.sub.i may comprise a photodiode, for example of the pn, pin or avalanche type, and an optical/electrical conversion device to convert the detected heterodyne signal into an electrical signal, for example an electrical voltage. The conversion device may comprise a filter to filter out the continuous component DC.sup.het of the detected heterodyne signal and keep only the alternating component AC.sup.heL; an amplifier of the amplitude of the filtered heterodyne signal, a detector and counter of the periodic variations of the filtered heterodyne signal in order thus to determine a value for the beat frequency of the heterodyne signal. The frequency value determined by each pixel P.sub.i may then be transmitted to a calculation unit which from this deduces a range map. The beat frequency can be determined simultaneously for all the pixels P.sub.i so as to obtain an instantaneous range map (range image) of the scene thus illuminated (this is an approach of the globalshutter type). As indicated above, the imaging system 1 may also be able to operate in rolling shutter mode and thus offer sequential reading of the pixels P.sub.i, for example row by row, where all the pixels P.sub.i of the one same row are read simultaneously.
[0095] Thus, in operation, the laser source emits the coherent primary signal, which is continuous and frequency modulated (laser radiation L), and of which part (object beam L.sub.o) is projected onto the scene to illuminate it instantaneously. Part (reference beam L.sub.ref) of the primary signal is transmitted to the detector array Det. The optical imager Im receives the reflected object beam L.sub.or and forms the image of the illuminated scene on the array of pixels P.sub.i. In this instance, it also receives the reference beam L.sub.ref, which is transmitted toward the array of pixels P.sub.i. The reference beam L.sub.ref and the reflected object beam L.sub.or interfere and form, in each pixel P.sub.i, a recombined beam L.sub.rec, which is to say the heterodyne signal. Each of the pixels P.sub.i advantageously determines the beat frequency f.sub.R of the detected heterodyne signal and the range map of the illuminated scene is thus determined.
[0096]
[0097] In the example of
[0098] An optical element 21, in this instance a transmission diffuser, is here positioned downstream of the splitter cube 31, and is designed to make the object beam L.sub.o divergent. The angular aperture of the object beam L.sub.o is in this instance such that the entire scene is illuminated uniformly by the object beam L.sub.o. The intensity profile of the object beam L.sub.o changes from a gaussian profile to a flat profile. In addition, the same diffuser 21 here renders the object beam L.sub.o divergent. A lens (not depicted) may be situated downstream of the diffuser (or upstream thereof) so that the angular aperture has a predefined value, in this instance a value substantially equal to the field angle of the optical imager Im. The optical recombiner SR here comprises a combiner cube 34 providing transmission of the reference beam L.sub.ref and of the reflected object beam L.sub.or toward the detector Det along the same optical axis. The optical imager Im here comprises a collection of lenses 41, 43 and an aperture diaphragm 42 (which defines the physical pupil) which are designed to create the image of the illuminated scene on the reception plane of the detector Det.
[0099] Furthermore, the optical splitter/recombiner SSR here comprises a diffuser 32 and a lens 33 which are situated between the splitter cube 31 and the combiner cube 34. The diffuser 32 is able to make the angular distribution of the intensity constant in order thus to obtain a flat profile for the intensity of the reference beam L.sub.ref rather than a gaussian profile. The diffuser 32 and the lens 33 shape the reference beam L.sub.ref to a predefined angular aperture, bearing in mind the various optical elements present, so that the reference beam L.sub.ref effectively illuminates all the pixels of the detector Det. Such a lens 33 is not indispensable for that. Note too that the diffuser 32 and the detection plane are not conjugate, so that the image of the diffuser 32 is not formed on the detector Det. The optical splitter/recombiner SSR is also configured to form an image, referred to as intermediate image PS, of the reference beam in a plane perpendicular to said optical axis, referred to as intermediate image plane PI. In other words, the diffuser 32 is situated in a plane conjugate with the intermediate image plane PI, to form the intermediate image PS (equivalent light source), which is situated in the segment [MM′] (and therefore in the central zone Zc) described later.
[0100] In the example of
[0101] Furthermore, the inventors have found that the performance of the imaging system D is optimized when the presence of interference fringes produced by the coherent superposition of the reflected object beam L.sub.or with the reference beam L.sub.ref in each pixel P.sub.i is limited or even avoided. Specifically, such fringes may lead to attenuation of the resultant oscillations of the beating of the heterodyne signal, insofar as the pixels P.sub.i integrate all the contributions indifferently. In order to limit or even avoid the appearance of these fringes, the imaging system D according to the invention is configured so that the wavefronts of the two incident portions L.sub.or,i and L.sub.ref,i are similar or identical in each of the pixels P.sub.i. This then makes it possible to limit the generation of fringes of interference between these two incident optical beams.
[0102] Also, according to the invention, the optical splitter/recombiner SSR is configured to form an image, referred to as intermediate image PS, of the reference beam L.sub.ref in a plane perpendicular to the optical axis AO, referred to as intermediate image plane PI. In addition, the position of the intermediate image PI on the optical axis AO is comprised in a segment [MM′] of the optical axis corresponding to the projection onto the optical axis of a zone of intersection (referred to as central zone Zc in
[0103] The equivalent light source PS has, at each point, a reference signal emission angle at least equal to the field angle of the optical imager Im so that each point of the equivalent light source PS (each source point) illuminates all the pixels P.sub.i. Moreover, the equivalent light source PS may advantageously have a lateral dimension at least equal to that of the central zone Zc of the rays of the full-light edge-of-field back-scattered object signal.
[0104]
[0105] So, the optical splitter/recombiner SSR creates the image of the reference beam L.sub.ref in the intermediate plane PI. The diffuser 32 (
[0106] The wavefront of the reference beam L.sub.ref incident on each pixel P.sub.i is thus made closer, or even identical, to that of the reflected object beam L.sub.or in terms of shape and main direction. The generation of fringes of interference between these two optical beams is thus limited, such interference fringes being liable to lead to degradation of the quality of the detected heterodyne signal. In other words, the array of pixels P.sub.i of the detector Det acts as a field diaphragm.
[0107] The field of view (FOV, the angle of the field) of the optical imager Im is defined as being the angle in which the detector array Det is sensitive to the reflected object beam L.sub.or reflected through the optical imager Im. It is defined here as being the angle, measured at the center of the exit pupil, between two incident rays of the reflected object beam L.sub.or passing through the one same centerpoint of the output pupil and arriving on two end pixels opposite to one another with respect to the optical axis AO. An end pixel is a pixel P.sub.i situated at the edge of the array of pixels. Furthermore, the exit pupil is the image of the aperture diaphragm from lenses situated downstream of the latter. In this instance, the centerpoint is situated in the plane of the aperture diaphragm.
[0108] Next, as illustrated more specifically in
[0109]
[0110] What is depicted is, for the two end pixels Px.sub.a and Px.sub.b, on the one hand, the full-light edge-of-field rays of the reflected object beam L.sub.or (in solid line), and, on the other hand, the rays coming from two opposite ends of the equivalent light source PS and received by each of the end pixels (in dotted line).
[0111] Thus, as far as the end pixel Px.sub.a is concerned, it here receives the two rays L.sub.or1.sup.a, L.sub.or2.sup.a of the reflected object beam L.sub.or passing through the edge of the aperture diaphragm Diaph (in the longitudinal plane) and the two rays L.sub.ref1.sup.a, L.sub.ref2.sup.a of the reference beam L.sub.ref coming from the edge of the equivalent light source PS (and of course, all the rays L.sub.or that fall between the rays L.sub.or1.sup.a, L.sub.or2.sup.a, and all the rays L.sub.ref that fall between the rays L.sub.ref1.sup.a, L.sub.ref2.sup.a). The rays L.sub.ref1.sup.a and L.sub.ref1.sup.a both pass through the same edge of the aperture diaphragm Diaph and are therefore superposed. Regarding the rays received by the second end pixel Px.sub.b, these are not detailed given that their situation is similar.
[0112] The optical splitter/recombiner SSR is designed so that the equivalent light source PS has, at each of its points, a divergence (or equivalent emission) cone that covers the entire array of pixels P.sub.i. Also, each of the pixels P.sub.i effectively receives the reference beam L.sub.ref in addition to the reflected object beam L.sub.or. The heterodyne signal is therefore indeed formed on each pixel P.sub.i.
[0113] Because the equivalent light source PS is positioned between the points M and M′ and has a lateral dimension at least equal to that of the central zone Zc, the reception cone for the reference beam L.sub.ref is substantially identical to the reception cone for the reflected object beam L.sub.or, and is so at each pixel P.sub.i.
[0114]
[0115]
[0116] It can be seen here that the difference between the wavefronts of these optical beams has been reduced by comparison with the scenario in which the equivalent light source PS is situated outside of the central zone Zc: in particular, the value of the angular deviation θ.sub.a between the main directions of these wavefronts which are incident on the pixel Px.sub.a has been reduced. A main direction is defined as being the mean direction of the angular cone of the incident beam (reference beam L.sub.ref or reflected object beam L.sub.or) at the pixel P.sub.i concerned, in this instance at the end pixel Px.sub.a.
[0117] Thus, forming an equivalent light source PS of the reference beam L.sub.ref in the central zone Zc makes it possible to reduce the angular deviation θ.sub.i at the pixel P.sub.i concerned. The dimension of the fringe spacing of the fringes that may form between the reflected object beam L.sub.or and the reference beam L.sub.ref is thus increased. The dimension of the fringe spacing is, in the first order, of the order of λ/2 sin θ.sub.i, where λ is the wavelength of the incident optical signals. Having angular cones for the optical beams (reference beam L.sub.ref and reflected object beam L.sub.or) incident on each pixel P.sub.i which are very similar or even identical to one another also makes it possible to increase the dimension of the fringe spacing.
[0118] In other words, it is advantageous, at each pixel P.sub.i, for an axis of propagation of the portion L.sub.or,i of the reflected object beam L.sub.or to be colinear or substantially colinear with an axis of propagation of the portion L.sub.ref, of the reference beam L.sub.ref. In instances in which the reflected object beam L.sub.or and reference beam L.sub.ref are approximately flat waves at each pixel P.sub.i and if the directions of propagation of these two waves form an angle θ, the fringe spacing on the sensor thus being
the condition for having at least one fringe spacing per pixel (flat tint) is expressed as
where a.sub.pix is the pixel size, for example θ<7° for λ=1300 nm and a.sub.pix=5 μm. As indicated further on, this then tends to yield “flat tint” interference fringes.
[0119] The detected intensity of the heterodyne signal is thus improved. Specifically, a significant angular deviation θ.sub.i between the main directions of the wavefronts and/or a deviation between the angular cones of the incident optical beams lead to a reduction in the size of the fringe spacings, which can then be of the order of size d.sub.px of the pixels P.sub.i, or even less. Also, during the photodetector acquisition time, the fact that several interference fringes (bright zones and dark zones) are present at the scale of each pixel P.sub.i may manifest itself in a degradation of the quality of the detection of the heterodyne signal.
[0120] It is particularly advantageous for the equivalent light source PS to be situated as close as possible to the aperture diaphragm Diaph, and for its width to be at least equal to and preferably substantially equal to that of the central zone Zc, and therefore in this instance to that of the aperture diaphragm Diaph.
[0121] In this respect,
[0122] A processor UT is connected to the detector and to the laser source in order to determine the range of the points in the scene which have been imagined on the pixels and in order to construct an instantaneous range image of the scene. This scene construction is performed using the calculated beat frequency F.sub.i associated with each pixel P.sub.i and from the optical frequency modulation of the laser radiation. What is meant here by range image is a map of the range of the various points in the scene observed each corresponding to a different pixel.
[0123] The imaging system D may thus determine an instantaneous (or “global shutter”) range image in which all the points in the scene are measured simultaneously with a measurement start and end that are identical for all the pixels (provided that the detector Det permits this). Furthermore, the imaging system D does not require “beam steering” means to scan the scene at a high rate. For operation at video speeds, the measurement duration may extend to 20 ms, which releases constraints both on the speed of the imager and the optical frequency modulation to be provided by the laser source. With all other parameters being equal, at video speed, what is obtained on the object path is the same photon flux on each of the pixels of the FMCW imager as on the photodetector of each pixel of the imaging system D (see
[0124] Thus, the invention proposes an active range imaging system D with high spatial resolution (number of points in the image as given by the number of pixels) which is robust in relation to parasitic ambient light, possibly of the “global shutter” type, and capable of operating at video speeds.
[0125] According to an embodiment, the imaging system D comprises a bandpass optical filter positioned in front of the detector or in front of each pixel of the detector intended for detecting range, for example in the case of the acquisition of images of RGB-z type and centered on the emission wavelength of the laser so as to filter out a high proportion of the parasitic ambient light.
[0126] The equivalent light source (intermediate image) PS of the reference beam L.sub.ref is produced in the intermediate plane PI in a number of different variant ways. In a first variant, the equivalent light source PS is produced using diffraction gratings in an integrated optics circuit contained in the optical recombiner and generating a plurality of “source points” in the intermediate plane PI. In this embodiment, the equivalent light source PS is therefore real. In a second variant, the equivalent light source PS is produced using an intermediate optical device, in free space, creating the intermediate image of the reference beam L.sub.ref. The equivalent light source PS here is the image of the reference beam L.sub.ref focused by an image lens or group of lenses. In this embodiment, the equivalent light source PS is generally virtual.
[0127]
[0128] The projection optical device Proj may be designed to project the light onto the scene that is to be illuminated of which the image is thereafter to be formed on the detector, for example a scene of rectangular shape. As a preference, the projection optical device Proj illuminates the scene in a cone of angular aperture substantially equal to the field of view of the optical imager Im (this being determined by its focal distance and the size of the imager). Thus, whatever the range of the scene, its image corresponds to the size of the imager.
[0129] In addition, the projection optical device Proj is preferably designed to illuminate the scene uniformly, in order thereafter to provide uniform illumination and a uniform signal-to-noise ratio on the detector, particularly in instances in which the scene is Lambertian. What is meant here by substantially uniform is that the standard deviation of the light intensity on the scene is, for example, less than 25% of the mean of the light intensity. According to one embodiment, the optical projection device Proj comprises a conventional two-stage optical device made up of a collimating first lens LP.sub.1, and of a second lens LP.sub.2 which creates the image of the first lens on the scene (for example, the first lens is approximately at the object focal point of the second lens LP.sub.2 in order to prevent an image of the source appearing on the scene which is practically at infinity). The laser source SL is the object focal point of the collimating lens LP.sub.1. In this embodiment, the set of lenses LP.sub.1, LP.sub.2 is folded to reduce the vertical size of the optical projection device Proj and therefore comprises two mirrors Mi.sub.1, and Mi.sub.2 positioned at 45°.
[0130] In this embodiment, the imaging system D further comprises an optical shaping device DOE (a Diffractive Optical Element) containing periodic patterns of which the period is of the order of the wavelength of the laser radiation. The optical device DOE is positioned after the optical projection device Proj in order to improve the uniformity of the illumination of the scene. Alternatively, the optical device DOE may be omitted and the shaping of the beam may be performed by the optical projection device Proj. The optical projection device Proj and the optical shaper DOE may be used alone or in combination in all the embodiments of the invention.
[0131]
[0132] The reference grating RR here is a grating of the High Contrast Grating (HCG) type allowing the decoupling of the light preferably on just one side of the waveguide (in
[0133] According to one embodiment, the waveguide and the grating RR of the second integrated optical circuit OI.sub.2 are produced on a surface of the or of one of the lenses of the optical imager Im. The losses through diffraction of the flux are therefore small insofar as the waveguide is small in lateral dimension and is not imaged on the detector Det, and the surface area of the grating RR is small in relation to the surface area of the pupil.
[0134] In an embodiment variant, the reference beam L.sub.ref is recombined with the reflected object beam L.sub.or by a plurality of reference diffraction gratings. There is therefore a relationship concerning the phase of the reference gratings to make them suitable for creating a wavefront on the pixels of the detector Det whereby the tangent at each point makes an angle no greater than
with respect to the tangent of the wavefront of the reflected object beam L.sub.or. For that, the guides of the circuit OI.sub.2 are advantageously single-mode waveguides. According to one embodiment, the circuit OI.sub.2 comprises active or inactive phase shifts in the waveguides or at the exit of the reference grating. Each diffraction grating behaves like a secondary source. The gratings are arranged uniformly to form an array of rows and columns, preferably with the same period. As a preference, the light diffracted by the collection of gratings forms light patterns uniformly spaced in the plane of the detector Det, with one pattern per pixel. The advantage of having a plurality of reference gratings is the ability to better distribute the lighting patterns, better control the uniformity of the illumination of the detector by the secondary source points and of better adjusting the wavefront of the reference beam L.sub.ref with respect to the wavefront of the reflected object beam L.sub.or on the detector Det.
[0135] Alternatively, according to another embodiment, the guided reference beam L.sub.rer is decoupled in the direction of free space toward the optical imager Im using a semireflective blade comprised in the waveguide and oriented in this instance at 45° to the axis of propagation in the waveguide. The advantage of the grating embodiment is that it reduces flux losses on the reference path by comparison with the semireflective blade embodiment.
[0136] In this example, the optical recombiner SR makes it possible to reduce the total thickness of the imaging system D because a bulky microoptics component (for example a semireflective blade of which the vertical size is approximately equal to the diameter of the entry pupil of the optical imager Im) is replaced with a simple substrate of thickness close to or less than one millimeter.
[0137]
[0138]
[0139] This embodiment has the advantage of avoiding the flux losses brought about by the coupling/decoupling of the reference beam L.sub.ref in the optical fiber FO of the imaging system D of
[0140] In the embodiments of
[0141]
[0142] Alternatively, according to another embodiment of this second variant, the assembly formed by the optical projection device Proj and the optical shaping element DOE is positioned downstream of the optical splitter LS on the path of the object beam L.sub.o.
[0143]
[0144] Alternatively, according to another embodiment, the laser source SL is configured to emit vertically, in order to save even more horizontal space. The first mirror Mi.sub.1 of the folded optical projection device Proj is therefore omitted and the laser source SL emits the laser radiation vertically toward the optical splitter LS through the collimating lens LP.sub.1. This embodiment is compatible with a longer focal length because the thickness of the optical imager Im is not limited by the width of the imaging system D. Thus, the angular resolution is greater, giving access to even more details in the image (although with a more limited field of view).
[0145] The imaging system D uses the FMCW technique and the laser source therefore operates in continuous wave emission. The laser source SL here can be configured so that the optical frequency is modulated by a periodically linear ramp of excursion B and duration T. In this embodiment, the detector is an array of pixels specializing in FMCW heterodyne detection with photodetection and signal processing integrated into each pixel, providing the beat frequency at the end of the frame time (20 ms in the case of a video speed frame time).
[0146] As illustrated in
[0147] The interconnection layer BE is positioned on one side of the photodetector Ph which is the opposite side to a side detecting said portion L.sub.or,i of the reflected object beam L.sub.or and the portion L.sub.ref,i of the reference beam L.sub.ref (BSI, or back-side illumination, configuration). The electronic circuit associated with the pixel P.sub.i is configured to collect the photoelectron signal generated in the photodetector Ph and indicative of the beating, and convert it into a voltage and then to: [0148] substitute a DC component of the heterodyne signal representative of the beat frequency with a predetermined fixed voltage value (and the mean value of the dark current), and generate an electrical signal referred to as substituted signal; then [0149] amplify said substituted signal, thus generating an amplified signal; [0150] detect, using a comparator, peaks or rising fronts or falling fronts of the amplified signal; [0151] proceed to count Sig the number N of detected periods T.sub.e in the amplified signal.
[0152] Various peak detection techniques can be implemented such as, for example, for example making a comparison against a predetermined level or comparing a sample against its two closest neighbors, this method being more robust toward low frequency noise.
[0153] The metallic interconnections transmit to the processor a signal representative of the count of the number of periods detected. The processor is therefore configured to determine the range z of a point in the scene, which point is associated with the pixel P.sub.i, by calculating the number N of periods T.sub.e detected during the duration T of the modulation of the optical frequency on the basis of the signal indicative of the count. Specifically, as explained previously, for a linear ramp, the frequency of the oscillations is
The number N (N≈Tf.sub.R, where f.sub.R=1/T.sub.e) of periods measured during the duration T makes it possible to calculate the range of a point in the scene z=Nc/2B. The resolution in terms of range is
[0154] The pixel P.sub.i thus detects the sinusoidal heterodyne signal indicative of the beating of which the duration is the frame time of the sinusoid of highest frequency F.sub.R,max corresponding to the frequency of the beating associated with an object in the scene situated at the maximum predetermined range z.sub.max, such that: F.sub.R,max=.sup.2Bz.sub.max/cT where c is the speed of light in a vacuum.
[0155] In order to avoid undersampling, according to Shannon's theorem, the sampling frequency for the pixel P.sub.i is greater than two times, preferably three times the frequency F.sub.R,max. The optical coupling in the photodetector present in each pixel P.sub.i is essentially vertical, which means to say that the light from the reference and object paths propagates in the thickness of the photodetective layer CP.
[0156] The streams of data exiting the detector Det may be similar to that of a standard image sensor operating at video speeds, the illumination information being replaced by the frequency information.
[0157] According to one embodiment, the light incident on the pixel P.sub.i is focused on the photodetector by a microlens ML if the fill factor of the pixel P.sub.i is less than 1. Such a factor may be desirable in order to avoid crosstalk (spreading of the light) between adjacent pixels of the detector Det. In that case, elements are arranged in the photodetective layer CP, between each pixel P.sub.i, in order to individualize zones in the photodetective layer. The microlenses may be contiguous, attached to a surface of the array of pixels and designed to optimize a collection of the reflected object beam L.sub.or and of the reference beam L.sub.ref on the pixels.
[0158] According to one embodiment, the laser source SL is an edge emitting laser (EEL) or a vertical-cavity surface-emitting laser (VCSEL) and the photodetector is a photodiode or an avalanche photodiode. Alternatively, in another embodiment, the laser source is a quantum cascade laser (QCL) and the photodetector is a bolometer array or an infrared photodetector of the mercury cadmium telluride (MCT) type. EEL or VCSEL lasers are preferably used for laser emission at wavelengths shorter than 3 μm, and QCLs are preferably used for laser emission at wavelengths greater than 3 μm.
[0159] The table which follows indicates, nonlimitingly, various technologies suitable for the creation of the imaging system D of the various embodiments of the invention:
TABLE-US-00001 Wavelength 850 nm, 1.3-1.55 μm 3-5 μm 8-14 μm 940 nm Laser source SL AsGa EEL InP EEL QCL QCL Free space optics Glass/plastic Glass Ge or Si Ge or Si Integrated optics SiN or glass Si or SiO.sub.2 Si or SiO.sub.2 Ge or Si Optical filter aSi/SiO.sub.2 aSi/SiO.sub.2 aSi/SiO.sub.2 aSi/SiO.sub.2 bandpass bandpass bandpass bandpass Photodetector Ph Si photodiode InGaAs Cooled MCT Micro- or Ge or avalanche bolometer photodiode photodiode
[0160] What is known by free space optics is the propagation of an optical beam through splitter cubes and blades and lenses. What is meant by integrated optics is the propagation of an optical beam through waveguides and diffraction gratings of integrated optical circuits.
[0161]
[0162] Consider two pixels P.sub.1 and P.sub.2 on the detector Det at the edge of the field. These are end pixels, insofar as they are situated at the edge of the array of pixels. The sub-beams SF.sub.1 and SF.sub.2 are defined as being the edge-of-field beams of the reflected object beam L.sub.or respectively illuminating the pixel P.sub.1 and the pixel P.sub.2 of the detector. The optical splitter/recombiner SSR is configured so that the longitudinal position (which is to say the position of the intermediate image plane PI) and lateral position of the intermediate image PS with respect to the optical axis cause the direction of the rays of the reference beam L.sub.ref to tend as far as possible toward that of the rays of the reflected object beam L.sub.or.
[0163] The subbeam SF.sub.1 (also referred to as L.sub.or,1) is the portion of the reflected object beam L.sub.or that illuminates the pixel P1. This is a subbeam that is parallel and inclined with respect to the optical axis, coming from a source point of the scene considered to be situated for example at infinity from the point of view of the imaging optical device Im.
[0164] The subbeam L.sub.ref,1 is the portion of the reference beam L.sub.ref that reaches the pixel P.sub.1 concerned. Each ray of this subbeam comes from a different source point. In the object space of the imaging optical device Im, these rays are parallel to the inclined subbeam of the reflected object beam L.sub.or because they pass through the same optical imager Im.
[0165] Also, the longitudinal position of the intermediate image PS on the optical axis AO is therefore comprised in the segment [MM′]. The segment [MM′] was defined above in connection with
[0166] Source points of the intermediate image PS (equivalent light source) placed in the intermediate plane PI but beyond, on either side of the segment [S1S2] would only partially illuminate the detector Det. Likewise, source points of an intermediate image situated too far upstream of the point M on the optical axis AO would illuminate the detector Det only partially.
[0167] The lateral dimension of the intermediate image PS from the point of view of the photometry (to limit losses on the reference path) is advantageously less than or equal to the lateral dimension of the intersection of the subbeams SF.sub.1, SF.sub.2 at the intermediate image plane PI. An advantageous dimension is represented by the segment [S1S2].
[0168] The overlapping of the angular spectra of the reflected object beam L.sub.or and of the reference beam L.sub.ref on a given pixel is greater if the aperture AP of the optical imager Im is small with respect to its field of view FOV on the reflected object path, this field of view angle being equal to the divergence of the reflected object beam L.sub.or. These two angular spectra are centered for the pixel of the detector Det situated on the optical axis AO.
[0169] If the intermediate image plane PI is situated in the plane of the thin lens with a diameter equal to the pupil, the wavefronts of the reflected object beam L.sub.or and the reference beam L.sub.ref are coincident. As indicated above, the quality of the detected heterodyne signal is thus greatly improved.
[0170] According to one embodiment of the invention, the optical splitter/recombiner SSR is configured so that a divergence of the reference beam L.sub.ref coming from the optical splitter/recombiner SSR is equal to the angle formed by said subbeams SF.sub.1, SF.sub.2 at the intermediate image plane PI. Thus, the reference beam L.sub.ref coming from the optical splitter/recombiner SSR illuminates the entirety of the pixels of the detector Det. The optical recombiner SR is configured so that the reference grating or gratings form source points with a divergence equal to the field of view of the reflected object beam L.sub.or at the intermediate image plane PI.
[0171] According to the invention, the optical splitter/recombiner SSR is configured so that the intermediate image plane PI is situated near to or coincident with a plane comprising a pupil or an aperture diaphragm Diaph of the optical imaging device Im, so as to improve the uniformity of the illumination on the detector, and also improve the quality of the detected heterodyne signal.
[0172] According to one embodiment, the intermediate plane PI is coincident with a plane containing a pupil or aperture diaphragm Diaph of the optical imager Im. This embodiment makes it possible to minimize the mean angle of the wavefront of the reflected object beam L.sub.or and that of the wavefront of the reference beam L.sub.ref on each pixel. Thus, the angular deviation between the reflected object beam L.sub.or and the reference beam L.sub.ref on each pixel is low enough to generate flat-tint fringes. What is meant here by “flat tint” is that there is no visible fringe on the pixels. In other words, the fringe spacing is large in comparison with the size of the pixels (in this instance the size of the microlenses).
[0173] The intermediate image plane PI may, however, be situated upstream of the optical imager Im and the pupil thereof. This position gives greater accessibility to the intermediate image PI and this may be advantageous in the case of an imaging system D in guided optics configuration (cf.
[0174] The intermediate image plane PI may also be situated downstream of the pupil of the optical imager Im, while at the same time keeping at least one lens of the optical imager Im that is common to the reflected object beam L.sub.or and the reference beam L.sub.ref so as to generate an equivalent light source PI (which in this instance is virtual).
[0175]
[0176] The two subbeams SF.sub.1 and SF.sub.2 (L.sub.or,1, L.sub.or,2) here are the full-light edge-of-field beams of the reflected object beam L.sub.or, respectively illuminating the pixels P.sub.1 and P.sub.2 of the detector Det. The lateral dimension of the intermediate image PI is given by the segment [S1S2], which is to say the lateral dimension of the intersection of the subbeams SF.sub.1, SF.sub.2 at the intermediate image plane PI. Emission cones with similar divergence for three source points are illustrated by way of example. Also depicted is the subbeam L.sub.ref,1, which is the portion of the reference beam L.sub.ref that reaches the pixel P.sub.1 concerned. As mentioned previously, the configuration of
[0177]
[0178] Thus, if several speckles are present in the one same pixel, the photodiode which indifferently integrates all the contributions will supply an averaged signal having an AC component that is reduced, or even eliminated (for example in the case of two speckles in phase opposition: cf.
[0179] Also, in one embodiment of the invention, the surface area (or size) of the photodetector is less than the surface area (or mean size) of the speckles. The mean size ϕ.sub.g of the speckle in the image of a diffusing object has a statistical mean of 2×λ×f.sub.#, where λ is the wavelength of the laser source, and f.sub.# is the aperture number of the optical imager Im. Thus, in this embodiment of the invention, the numerical aperture of the optical imager Im is designed so that the mean size ϕ.sub.g of the speckles on the detector Det is greater than a characteristic dimension a.sub.pix of the photodetector of each pixel of the detector. What is meant by characteristic dimension is, for example, the side length of a square photodetector (in this instance the side length of the associated microlens).
[0180] The following passage describes the parameters and typical performance of the imaging system D according to one embodiment of the invention. The parameter value used is nonlimiting and given merely by way of example.
[0181] In this example, it is assumed that the predetermined maximum range of the scene with respect to the detector Det is z=10 m. It is assumed that the desired resolution in terms of range is 5 cm, and this fixes the excursion of the optical frequency of the laser as
[0182] In order to operate at video speeds, it is desired to create a range image of the scene in 1/50 Hz=20 ms at most. The duration T of the optical frequency ramp of the source is therefore 20 ms. Alternatively, the duration T of the ramp may be a submultiple of 20 ms, in order to be able to perform several ramps in this duration T, establish measurement statistics and thus improve the precision of the measurements. However, doing that is possible only if the signal and the signal-to-noise ratio are sufficient.
[0183] The gradient of the ramp is 0.15 MHz/μs, which is a value readily accessible by modulating the injection current of a semiconductor laser. The frequency of the oscillations that will be detected by the pixels of the detector is
the period of these oscillations 1/f.sub.R≈100 μs.
[0184] In the case of a field of view FOV of 26×19° (32° on the diagonal) and an optical imager Im with a focal length f=7 mm, the minimum size of the detector Det is therefore equal to 2×f×tan(FOV/2)=3.2×2.4 mm. For a VGA (640×480 pixels) format, the pixel size is therefore 5×5 μm.
[0185] In the case of a laser wavelength A of 1300 nm (with, for example, a germanium or InGaAs detector), the aperture number f.sub.# of the optical imager Im is preferably greater than 1.9 in order thus to have a characteristic speckle size ϕ.sub.g in the image plane that is greater than the pixel size a.sub.pix of the detector Det. What is selected for example is f.sub.#=2 with a pupil Diaph of the optical imager Im of 3.5 mm.
[0186] The power of the laser of the object beam L.sub.o is equal for example to 10 mW, a value that meets retina safe standards at 1300 nm. A ratio T.sub.s of 95%/5% between the emitted powers of the object/reference paths is chosen, for example, giving a power of 0.5 mW on the reference beam L.sub.ref.
[0187] Considering a 20% Lambertian diffuse reflectance factor for the scene Obj and an optical recombiner which is a semireflective cube with a T.sub.LO ratio equal to 95%/5% for recombination, the fraction of the power of the portion L.sub.or,i of the reflected object beam that is incident upon each pixel P.sub.i with respect to the object power emitted is 10.sup.−15. This ratio is extremely small despite the relatively large aperture of the optical imager Im because a large proportion of the light of the object path is diffused outside of the pupil of this optic, the portion passing through the pupil then being shared between all the pixels P.sub.i of the detector Det. The fraction of the power of the reference path that is incident in each pixel is markedly higher, here of the order of 10.sup.−6.
[0188] It is now assumed that the photodetector Ph is an InGaAs photodiode with a quantum efficiency QE of 70% at 1300 nm. In addition, in order to conform to Shannon's theorem, the integration time (the inverse of the sampling frequency) is equal to one third of a period 1/f.sub.R. The next equation allows the photon flux F on one pixel P.sub.i of the detector to be calculated:
where R is the reflectance of the Lambertian scene, and T.sub.s, T.sub.LO and T.sub.opt are the transmission of the optical splitter LS, of the optical recombiner SR and of the optical imager Im respectively and ϕ.sub.opt is the diameter of the pupil of the optical imager Im, N.sub.pix is the number of pixels of the detector, and E.sub.sun is the illumination from the sun.
[0189] Following detection of the heterodyne signal, the continuous component DC of the detected signal (which is to say the electronic signal of the mean flux of the object path and of the reference path incident on the pixel) obtained during the integration time is F.sub.DC=13000 electrons. The alternating component AC, i.e. the useful signal F.sub.AC, (the electronic signal indicative of the beating) modulated at the frequency f.sub.Rmax, carrying the range information varies between 15 and 140 electrons depending on the (unpredictable) positioning of the integration times within the sinusoidal signal.
[0190] The dominant noise is the noise of photons on the reference path at each of the pixels, corresponding to approximately 100 electrons (equal to √{square root over (F.sub.DC)}) in the integration time. The noise of photons in the ambient parasitic light F.sub.amb is smaller (20 electrons of noise with sunlight filtered through a bandpass filter of width 50 nm centered on the wavelength of the laser). The dark current noise (approximately 10 electrons with a dark current of 1 pA) and the reading noise (a few electrons) are likewise far lower than the photon noise. This being so, the signal-to-noise ratio, of the order of 1 or even below, is compatible with the detection and counting of the peaks in the useful signal F.sub.AC.
[0191] In the case of a single-point FMCW LIDAR imaging system which scans a scene (angular sweep) with VGA resolution at video speed, having the same parameters as those used in the previous calculation, the signal-to-noise ratio is similar. However, the detector would need to have a pass band of at least 10 GHz, which is possible but restrictive, and the laser source would need to be able to be wavelength modulated at a frequency of several tens of MHz with a modulation amplitude of several GHz as explained previously. In addition, the illumination device would need to sweep the scene at a rate of several MHz. These values are not really compatible with present-day technology.