Banknote imaging

Abstract

A method of obtaining a plurality of infrared images of a banknote that involves simultaneously illuminating the banknote with infrared light at a first wavelength and infrared light at a second wavelength, capturing an image of the banknote with an RGB camera, obtaining from both a first output channel signal and a second output channel signal of the RGB camera sensor where the intensity distribution of the infrared light at the first wavelength and the intensity distribution of the infrared light at the second wavelength uses a first calibration coefficient and a second calibration coefficient of the RGB camera sensor, producing separate infrared images of the banknote at the first wavelength and the second wavelength from the respective intensity distributions.

Claims

1. A method of obtaining a plurality of infrared images of a banknote comprising: simultaneously illuminating the banknote with infrared light at a first wavelength and infrared light at a second wavelength, wherein the first and second wavelengths are different, and wherein the first wavelength is between 400 nm and 1000 nm, and wherein the second wavelength is between 400 nm and 1000 nm; capturing an image of the banknote with an RGB camera; obtaining from both a first output channel signal and a second output channel signal of the RGB camera sensor an intensity distribution of the light at the first wavelength and an intensity distribution of the light at the second wavelength using a first calibration coefficient and a second calibration coefficient of the RGB camera sensor; and producing separate infrared images of the banknote at the first wavelength and the second wavelength from the respective intensity distributions; wherein the first calibration coefficient is defined as a ratio of the first output channel signal amplitude at the light first wavelength to the second output channel signal amplitude at the light first wavelength; and wherein the second calibration coefficient is defined as a ratio of the second output channel signal amplitude at the light second wavelength to the first output channel signal amplitude at the light second wavelength; and wherein a first output channel includes a first output channel of the RGB camera sensor, and wherein a second output channel includes a second output channel of the RGB camera sensor, wherein the first and second output channels of the RGB camera sensor are different channels of the RGB camera sensor.

2. A method of claim 1, wherein the RGB camera sensor is a CMOS or CCD image sensor.

3. A method of claim 1, wherein the RGB camera sensor includes a Bayer filter matrix.

4. A method of claim 1, wherein the first wavelength is 730 nm and the second wavelength is 880 nm.

5. A method of claim 1, wherein the first and second wavelengths are emitted from a first infrared LED and a second infrared LED respectively.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) An embodiment of the present invention will now be described, by way of example only, with reference to the accompanying schematic drawings, in which:

(2) FIG. 1 shows a banknote validator optical arrangement configured for transmission mode.

(3) FIG. 2 shows the banknote validator optical arrangement configured for reflection mode.

(4) FIG. 3 shows wavelength distribution for the output channels of an RGB camera.

(5) FIG. 4 shows a detail of the wavelength distribution around the infrared wavelengths of interest.

(6) FIG. 5 is an image of a banknote in the visible region of the spectrum.

(7) FIG. 6 is an image of the banknote in the infrared region of the spectrum.

(8) FIG. 7 is an image of the banknote at a first infrared wavelength.

(9) FIG. 8 is an image of the banknote at a second infrared wavelength.

DETAILED DESCRIPTION

(10) As shown in FIG. 1, a banknote validator optical arrangement comprises a light source 1, an RGB camera 2 and a control unit 3. The banknote validator optical arrangement also includes an optional UV light source 5.

(11) In transmission mode, as shown in FIG. 1, the light source 1 illuminates a banknote 6 from below. The light source 1 includes a plurality of LEDs [not shown]. One LED operates in the visible region of the electromagnetic spectrum, typically a white LED, a second LED emits infrared light at a first wavelength, and a third LED emits infrared light at a second wavelength.

(12) In a preferred embodiment, the first wavelength is 730 nm, and the second wavelength is 880 nm. However, the reader will be aware that these wavelengths are examples only, and any wavelengths can be selected dependent upon the requirements of a given application.

(13) Visible light and infrared light 1′ passing through the banknote 6 under investigation is reflected by reflector element 4 towards the RGB camera 2.

(14) The RGB camera 2 is a conventional digital camera comprising optics and a sensor [not shown]. The sensor is typically a CCD sensor or a CMOS sensor. In either case, the sensor will include a conventional Bayer filter matrix. The output of the sensor is directly connected to the control unit 3 via a red channel 7, a green channel 8, and blue channel 9 in a known manner. The control unit 3 includes a processor [not shown] including RGB control logic for receiving, processing and analysing the RGB data received from the RGB camera 2.

(15) The optional UV light source 5, which includes at least one UV LED, illuminates the banknote with UV light 5′ which excites fluorescence in the banknote 6. The fluorescent image data from the banknote 6 is collected by the RGB camera and passed to the control unit 3 for analysis.

(16) In a preferred embodiment, three separate images of the banknote 6 are obtained in sequence: (i) image taken with white LED only on; (ii) image taken with infrared LEDs only on; and (iii) image taken with the UV LED only on.

(17) In a similar manner to that which is described above, and as shown in FIG. 2, the banknote validator optical arrangement can be configured for operation in reflection mode. Here, the RGB camera 2 is positioned to directly collect images of the banknote 6 without requiring a reflector element 4. However, the skilled reader will be aware that the banknote validator optical arrangement can be configured in such a manner that it may operate simultaneously in both reflection and transmission modes with the inclusion of a suitable two-way reflector element 4.

(18) As will be known to the skilled reader, a conventional Bayer filter matrix does not include any infrared filters. Consequently, to separate an infrared image from a visible image of a banknote it has been conventionally necessary to incorporate at least one infrared filter within the optical arrangement of the banknote validator. As noted above, this leads to a commensurately larger and more expensive optical arrangement.

(19) Advantageously, it has been found that an image of a banknote undergoing illumination by infrared light at two differing wavelengths can be separated into two distinct images by manipulation of the relative differences in the efficiencies of the colour filters within the Bayer filter matrix of a conventional RGB camera.

(20) FIG. 3 illustrates the characteristic outputs of the red, blue and green channels of an RGB camera sensor. Wavelength range 10 represents the area of the infrared spectrum that is of interest in the particular example discussed herein.

(21) FIG. 4 shows an exploded view of the wavelength range 10 shown in FIG. 3. The banknote 6 has been simultaneously illuminated by a first LED operating at a wavelength of 730 nm, and a second LED operating at a wavelength of 880 nm.

(22) The red channel trace 11 represents the spectral response of the red-pass filter of the Bayer filter matrix, and the blue channel trace 12 represents the spectral response of the blue-pass filter of the Bayer filter matrix.

(23) As shown in FIG. 4, the blue and red filters respond differently at 730 nm and 880 nm respectively. The coefficients α and β represent the relative amount of light transmission by the red-pass filter at 730 nm and 880 nm respectively. Likewise, the coefficients γ and δ represent the relative amount of light transmission by the blue-pass filter at 730 nm and 880 nm respectively.

(24) The coefficients α, β, δ and γ can be determined by a calibration process using calibration paper under separate illumination by the 730 nm LED and then the 880 nm LED. The use of calibration paper will be known to the skilled reader of the present application.

(25) Image Extraction

(26) To extract an image produced by the individual infrared LEDs it is necessary to quantify contributions from each LED to the red and blue filters by defining two ratios:

(27) $\begin{array}{cc}x=\frac{\gamma }{\alpha }& \left(1\right)\\ y=\frac{\delta }{\beta }& \left(2\right)\end{array}$

(28) For any given sensor pixel the intensity at wavelength 730 nm is given as I(a), and the intensity at wavelength 880 nm is given as I(b). So, for each pixel under a blue-pass filter, the total light intensity, B, is given by:
B=γI(a)+δI(b)  (3)

(29) Likewise, for each pixel under a red-pass filter, the total light intensity, R, is given by:
R=αI(a)+βI(b)  (4)

(30) From equations (1), (2) and (3) we obtain:
B=α×I(a)+βyI(b)  (5)

(31) From equation (4) for we obtain:
βyI(b)=yR−αyI(b)  (6)

(32) And from equations (5) and (6) we obtain:

(33) $\begin{array}{cc}I\left(a\right)=\frac{B-\mathrm{yR}}{\alpha \left(x-y\right)}& \left(7\right)\end{array}$

(34) Similarly, we obtain:

(35) $\begin{array}{cc}I\left(b\right)=\frac{B-\mathrm{xR}}{\beta \left(y-x\right)}& \left(8\right)\end{array}$

(36) Using equations (7) and (8), the RGB control logic calculates the image data for each of the infrared LEDs, which can then be combined to produce a distinct extracted image at 730 nm and at 880 nm respectively.

(37) FIG. 5 shows a banknote image 13 taken whilst illuminated by a white LED only. FIG. 6 shows an image 13′ of the same banknote as shown in FIG. 5 taken whilst illuminated by the 730 nm infrared LED and the 880 nm infrared LED.

(38) FIG. 7 shows a schematic representation of an extracted image for illumination by the 730 nm LED only. FIG. 8 shows an extracted image of the same banknote. Here, the image is what would have been observed if the banknote had been illuminated by the 880 nm LED only.

(39) In this way, two separate images of the banknote, FIGS. 7 and 8, can be extracted without the need for sequentially illuminating the banknote with the first infrared LED and then with the second infrared LED, or the need for incorporating infrared filters in the Bayer filter matrix of the RGB camera sensor.