System for acquiring a color image and an infrared image of a scene

Abstract

A system for acquiring a color image and an infrared image of a scene. The acquisition system (100) comprises: - an image acquisition device (1), sensitive to visible and infrared wavelengths, comprising: - a sensor (14); - a lens (10) configured to focus the light received at the input of the acquisition device (1) onto the sensor (14), a dual-bandpass filter (12) being placed in front of the sensor (14); - a processing module (16) configured to separate, at the output of the sensor (14), the infrared image (IIR) and the visible image (IVIS) and to define at least one signal to be controlled and an associated control setpoint for the infrared signal, according to which at least one artificial light source (20, 30) is controlled; - at least one artificial light source (20, 30) illuminating said scene, comprising an artificial infrared light source (20). Figure for the abbreviation: Fig. 1

Description

Description Title of the invention: System for acquiring a color image and of an infrared image of a scene

[0001] = The invention relates to a method for acquiring a color image and an in- infrared of a scene, as well as the acquisition system implementing said method.

[0002] …— The invention applies, in particular, to technical fields such as security for the recognition of biometric information, including faces, from a individual. State of the prior art

[0003] To realize a facial recognition system with anti-fraud, a system of camera capable of simultaneously acquiring both a near-in- infrared of the face, and an image in the visible spectrum is classically used.

[0004] — To reduce the cost and size of such a system, the document is known FR3102324 which describes a device for combining the two images on a single image sensor, this RGB image acquisition device does not include any infrared filter and including a high-pass filter configured to attenuate the blue component of light so as to recover infrared information on blue photosites. This device offers good resolution of the green component however in the light environments where such a device is used, there may be have very significant differences in level between infrared and visible, which makes it difficult and even impossible to exploit both infrared and visible images at the same time, due to the limited dynamic range of image sensors. In addition, with a single set of exposure parameter settings such as time exposure and gain, this device does not allow both to be correctly reproduced Images.

[0005] — Use cases in which the level differences are very significant between infrared and visible light concern in particular: - bright environments without natural ambient light, i.e. comprising ambient light only from an artificial source, such as a LED lighting from the acronym Light Emitting Diode, because These environments include almost no infrared light signals or near infrared, but only light signals in the visible; - bright environments where natural light is filtered by means of example of heat-resistant film placed on the windows, because these environments do not com- also take very little infrared light signal; - non-illuminated environments, because in complete darkness no signal is received. One solution may be to add visible illumination, also called white illumination, and infrared illumination to the device when the two cameras are separate, each being sensitive to a single illumination, either visible or infrared, but in the case of a single camera sensitive to both wavelengths, the direct application of this solution is not suitable. In addition, the integration of two cameras generates architectural, cost and space constraints. The present invention aims to at least partially overcome these drawbacks, possibly leading to other advantages. Statement of the invention The invention relates to a system for acquiring a color image and an infrared image of a scene, said acquisition system comprising: - an image acquisition device, said device being sensitive to visible and infrared wavelengths, comprising: - a sensor; - a lens configured to focus the light received at the input of the acquisition device onto the sensor, a dual-band pass filter being placed in front of the sensor, and in particular between the lens and said sensor; - a processing module connected to said image acquisition device configured to separate, at the sensor output, the infrared image and the visible image and to define at least one signal to be controlled and a related control instruction, said at least one control instruction being an infrared and / or visible signal control instruction according to which at least one artificial light source is controlled; - the at least one artificial light source illuminating said scene, said at least one source comprising an artificial infrared light source emitting a light signal in the infrared and / or an artificial visible light source emitting a light signal in the visible. This acquisition system allows, with a single camera, and therefore an advantage in terms of compactness, to obtain two images (infrared and visible) which can be used for biometric recognition or visual code reading, in all light conditions. For example, the dual-pass filter could consist of a network of filters. Advantageously, said image acquisition device comprises a dual bandwidth filter including: - in the case of a sensor with at least four different photosites, the first bandwidth of the filter is configured to allow visible wavelengths to pass through, in particular less than 650 nm, and in the case of a sensor with three different photosites, the first bandwidth of the filter is configured so as to attenuate the blue component of the light, in particular with a first bandwidth extending for example between 530 and 650 nm; and - the second bandwidth is configured to allow infrared wavelengths to pass, in particular the second bandwidth extends from 800 to 875 nm; which allows in particular, thanks to the restricted infrared band, to have repeatability in the rendering of infrared images versus external light conditions, the spectrum is therefore controlled. For example, in these sensors, the four different photosites include in particular the Red, Green, Blue and Infrared photosites and the three different photosites are in particular the Red Green and Blue photosites, knowing that according to the invention in the case of such a sensor with three photosites, this last photosite is no longer sensitive except to Infrared, due to the application of the filter, which allows good acquisition of faces since the blue component is hardly represented there. Advantageously, the second bandwidth extends from 760 nm-800 nm, which makes it possible to recover sufficient infrared signal even in the presence of an anti-heat film because the latter cuts around 800 nm. Indeed, without this widened bandwidth, in the case of an anti-heat film there could be more than 10klux of visible and almost no infrared, so the system would have to be able to provide infrared illumination capable of generating the same signal level as the visible illumination of 10klux to distinguish the infrared.But thanks to the bandwidth expansion, there is enough infrared to obtain an image, without the need for infrared illumination, the dimensioning of which can be reduced so that it is able to generate a signal level equivalent to a visible illumination of only 1000 lux, in order to respond to the use case corresponding to strong artificial ambient light (not emitted by the system), by LED for example. Advantageously: -the artificial source of light emitting in the infrared is configured to emit, in a range of distance of use, in particular from 40 to 100 cm, a signal equivalent to the signal generated by intense ambient visible light, in particular 1000 lux, for example an 850 nm light-emitting diode, the spectrum of which extends from 800 to 875 nm and / or -the artificial light source emitting in the visible range is configured to provide satisfactory lighting in complete darkness, in particular 50 lux at a usage distance of approximately 70 cm: which allows for the availability of visible and infrared or near infrared light sources. powerful infrared without presenting any risk or discomfort (glare) for the user, the infrared lighting can be modulated in intensity, notably between 0 and the equivalent of 1000 lux. Advantageously, the system includes another shooting device sensitive only in the visible, synchronized with the acquisition device and whose exposure parameters are controlled independently of those of the acquisition device, which allows stereoscopy and thus the detection of fraud of all types at limited cost. Advantageously, the other device can include a monochrome sensor or a sensor with at least Red, Green and Blue photosites, which makes it possible to obtain stereoscopy at reduced cost. The invention also relates to a method for acquiring a color image and an infrared image of a scene, said acquisition method comprising steps of: - receiving an image from an image acquisition device sensitive to visible and infrared wavelengths; - processing of said received image separating the infrared image and the visible image and defining at least one signal to be controlled and a related control instruction, said at least one control instruction being an infrared and / or visible signal control instruction according to which at least one artificial source of infrared and / or visible light is controlled: the process having the same advantages as the system. Advantageously, the received image includes metadata, such as gain and exposure time specific to said received image. Advantageously, the processing step comprises a step of detecting an object in the color and / or infrared image and determining the distance between the detected object and said acquisition device, which makes it possible to estimate the additional signal produced by the artificial lighting of the system on the object. Advantageously, the processing step comprises a step of calculating at least two metrics on all or part of the image, the first metric characterizing the infrared signal, the second metric characterizing the visible signal, which makes it possible to quantify the two types of signals and then to be able to compare them, in particular with each other. Advantageously, said part of the image is a region of interest of the image delimiting said detected object, said object being in particular a body part or information, such as a visual code, in particular printed on a physical document, which makes it possible to limit the calculation of the metrics to a zone of interest in the image, such as a face in the context, for example, of an application to face recognition. facial, or as information when applied to reading physical documents. Advantageously, the at least two metrics are calculated as being at least the average over all or said part of the image on each wavelength channel of said characterized signal, which makes it possible not to work with images close to saturation. Advantageously, the processing step includes: - a step of estimating a component of the first metric characterizing a quantity of infrared signal due to ambient infrared, i.e. not emitted by the at least one artificial source of infrared and / or visible light, and; - a step of calculating the distribution of ambient lighting, in particular in the form of the ratio of the quantity of the component of the first metric characterizing the infrared signal due to the ambient infrared to the second metric, and; - a step of comparing said distribution to a predetermined target distribution, and in that the control instruction of the infrared signal is a function of said comparison, and in particular in that the control instruction of the visible signal is a function of said comparison; which allows for a simple and resource-efficient calculation of light distribution. Advantageously, said predetermined target distribution is a function of the ratio of the infrared logic level to the visible logic level, said logic levels being determined in particular so that the sum of these two instructions is less than half of the maximum signal that can be restored by the camera. Indeed, knowing that on the camera, the visible channels are also sensitive to infrared, the sum of these two instructions is preferably less than half of the maximum logic level of the signal N',nax that can be restored by the camera, for example N,iax = 1023 in the case of a camera with a 10-bit output, so as to avoid saturations and maintain a sufficient margin. Advantageously, the control setpoint of the visible signal is a function of said comparison, in that if the calculated ambient lighting distribution is greater than the target distribution then the controlled signal is the infrared signal on the infrared setpoint and otherwise the controlled signal is the visible signal on the visible setpoint, which makes it possible to achieve the setpoint even in the case of lower ambient infrared, knowing that there are hardly any opposite use cases, i.e. in which there is a significant excess of infrared in the scene. Advantageously, the processing step includes: - a step of calculating the exposure parameter of the acquisition device to be applied, such as gain and / or exposure time, in particular as a function of the ratio of at least one control instruction on said related signal to be controlled, to achieve said at least one control instruction; and / or - a step of calculating the control of the artificial source of infrared and / or visible light; which makes it possible to achieve the control setpoint of the signal to be controlled using easily controllable parameters. Advantageously, said exposure parameters (in particular gain and exposure time) to be applied are in particular determined in such a way that their product is equal to the product of said parameters being applied multiplied by a ratio of the setpoint to be controlled by the controlled signal, which makes it possible to control in a proportional and linear manner. Advantageously, the step of calculating the control of the artificial infrared light source of the system is carried out by first calculating the additional infrared signal necessary to reach the control setpoint of the infrared signal knowing the ambient infrared signal, by subtracting from the control setpoint of the infrared signal said ratio multiplied by the ambient infrared signal: S, Radditionnel = Crr- Räti0 . Sppambiant * then calculating a duty cycle to be applied to the infrared lighting source. Advantageously, the control of the artificial infrared light source of the system is carried out by modulation of the duty cycle, which makes it possible to generate a signal equivalent to visible illumination between 0 and 1000 lux in the operating distance range. Advantageously, during the calculation step of the control of the artificial source of visible light of the system - If the white light was off, and the product of the exposure parameters exceeds a predetermined ignition threshold, then the white light is turned on; - If the white light was on, and the product of the exposure parameters is less than a predetermined extinction threshold, then the white light is turned off. Advantageously, the control for piloting the artificial visible light source of the system is binary, which allows simple control by hysteresis. Advantageously, the acquisition method comprises an initialization step carrying out an iterative scan of configurations until the object is detected, each configuration comprising at least one exposure parameter, such as gain and / or exposure time, and an intensity parameter of the at least one artificial light source, the values ​​of said parameters being fixed at predetermined values ​​in each configuration; which makes it possible to detect the desired object regardless of the lighting environment, because the fixed values ​​of the parameters of each configuration correspond to optimums in distinct use cases, and as many configurations as there are identified use cases are tested one after the other iteratively until the desired object is "caught". Advantageously, the predetermined target distribution is defined during the initialization step. Advantageously, said configurations are recorded in table form in the processing module. Advantageously, the acquisition method according to the invention is implemented by the processing module of the device according to the invention. The invention also relates to a method for authenticating a body part, characterized in that it comprises: - acquiring a color image and an infrared image of said body part by the acquisition method according to the invention and in that the detected object is a body part; and - authenticate the body part from said color and infrared images; The said authentication method has the same advantages as the acquisition method, and allows biometric recognition. The invention also relates to a device for authenticating a body part, comprising: - an image acquisition system according to the invention; and - the processing module being configured to detect a body part from the received image and to perform authentication of said body part detected from said color and infrared images; which has the same advantages as the acquisition system according to the invention and allows biometric authentication, in particular using the visible image to “recognize” the person and the infrared image to verify the authenticity of the face (in other words detect possible fraud). The invention also relates to a computer program, comprising instructions for implementing, by a device, the acquisition method according to the invention, when said program is executed by a calculation unit of said device, said program having the same advantages as previously mentioned. The calculation unit may in particular be that of the processing module of the device according to the invention. The invention further relates to storage means, storing a computer program comprising instructions for implementing, by a device, the acquisition method according to the invention, when said program is executed by a computing unit of said device, said means having the same advantages as previously mentioned. The computing unit may in particular be that of the processing module of the device according to the invention. Brief description of the drawings The invention will be better understood and its advantages will appear better on reading the detailed description which follows, given for information purposes and in no way limiting, with reference to the appended drawings in which: [Fig.1] schematically presents an image acquisition system according to one embodiment of the invention; [Fig.2] schematically illustrates an example of hardware architecture of a processing module of the image acquisition system according to one embodiment; [Fig.3] illustrates a method of acquiring an infrared image and a color image according to one embodiment: [Fig.4] schematically presents an image acquisition system according to another embodiment of the invention: [Fig.5] shows an authentication device according to one embodiment; and [Fig.6] illustrates an authentication method according to one embodiment. Identical elements shown in the above figures are identified by identical reference numerals. Detailed presentation [Fig.1] illustrates an image acquisition system according to one embodiment of the invention. The image acquisition system 100 comprises an image acquisition device 1 comprising an objective 10 which comprises at least one lens. The objective 10 is configured to focus the light received at the input of the system and more precisely of the acquisition device 1 onto a sensor 14, such as a CCD (Charge Couple Device) or CMOS (Complementary Metal Oxide Semiconductor) sensor. The sensor 14 is composed of a plurality of photosites. The image acquisition device 1 also comprises a dual-band pass filter 12 placed in front of the sensor 14 such that each photosite of said sensor sees only a single color, in particular between the objective 10 and said sensor 14. The dual-band pass filter 12 may in particular be glued to the objective or constitute a treatment layer on the objective or constitute a treatment of the glass of the objective 10, or even be positioned in front of the objective.Alternatively, the filter 12 could be a network of filters 0, notably placed at different positions in the optical path, for example a high-pass filter upstream of the lens and a network of color filters placed between the lens and the sensor. The sensor 14 may have only three different photosites or at least four different photosites. In the case of a sensor 14 comprising only three different photosites, the first bandwidth of the dual-band pass filter 12 is configured so as to attenuate the blue component of the light, so as to recover infrared information on photosites initially dedicated to blue so that each photosite of the sensor 14 sees only one color: the red channel is thus sensitive to Red and Infrared via a first photosite, the green channel is sensitive to Green and Infrared via a second photosite, and the infrared channel is sensitive to Infrared via a third photosite; the first bandwidth therefore extending for example between 530 and 650 nm. The device 10 obtained is then called RVPIR for Red Green Near Infrared. In the case of a sensor 14 comprising four different photosites, the first bandwidth of the dual-band pass filter 12 is configured to allow visible wavelengths to pass, such that each photosite of the sensor 14 sees only one color, the red channel is thus sensitive to Red and Infrared via a first photosite, the green channel is sensitive to Green and Infrared via a second photosite, the blue channel is sensitive to Blue and Infrared via a third photosite, and the near infrared channel PIR is sensitive to Infrared via a fourth photosite, the first bandwidth covering in particular wavelengths below 650 nm. Whatever the sensor 14, the second bandwidth of the filter 12 is configured to allow infrared wavelengths to pass, for example the second bandwidth extends from 790 to 875 nm, and preferably extends from 760 nm-800 nm, so as to recover sufficient infrared signal even in the presence of an anti-heat film, this type of film cutting the wavelengths around 800 nm, while only requiring a significant oversizing of the infrared lighting. The acquisition device does not include any infrared filter (called Infrared Cut), so it is sensitive to infrared on all its channels. The acquisition device 1 with a sensor 14 composed of only three photosites is obtained in particular from an acquisition device consisting for example of a conventional Red Green Blue RGB camera from which the infrared filter has been removed and in which the color filter network is replaced by a filter designed to only allow red, green and infrared IR to pass. The device 1 obtained is then called RVPIR for Red Green Near Infrared. The acquisition device 1 with a sensor 14 composed of at least four photosites is for example obtained from a Red Green Blue and near Infrared RGBPIR camera associated with a dual bandwidth filter, also called a filter notch, passing through the entire visible range, that is to say in particular between 400 nm and 650 nm, and passing through the infrared, that is to say in particular between 800 nm and 875 nm, and blocking the rest. The device 1 obtained is then called RGBPIR for Red Green Blue Near Infrared. The difference compared to the previous acquisition device 1 with a sensor 14 composed only of three photosites is that it lets the blue (between 400 nm and 530 nm) pass instead of cutting it. The acquisition device 1 also comprises a processing module 16. The processing module 16 receives as input a raw image Is from the sensor 14 and generates an infrared image I;r and a color image Ly;s. A raw image consists of the photonic information obtained, after the conversion by the photosites of the incident photons into a digitized electrical signal. In the case of a sensor 14 with only three photosites, the processing module 16 is configured to calibrate, by determining calibration parameters, in a known manner, the image acquisition device 1. Similarly, in the case of a sensor 14 with at least four photosites, the image device 1 is calibrated in a known manner. The processing module 16 is also configured to generate an infrared image Iyr and a color image 1yzs from a raw image lp acquired by the sensor 14 and the determined calibration parameters. The processing module 16 is therefore capable of implementing the method for generating the images Iyr and 1yzs. Advantageously, the processing module 16 generates images Ir and 1y;s which are synchronized. The acquisition device | may comprise other elements well known from conventional RGB cameras. These elements are not shown in Fig. |. The elements of the lens 10 may contain anti-reflection type optical treatments to increase the signal-to-noise ratio in the acquired images. The acquisition device 1 generally comprises a microlens array positioned above the sensor 14 whose role is to optimize the collection of light by the photosites. The acquisition device 1 also comprises a signal processing module (not shown in Fig. 1) at the output of the sensor 14. Its role is to reconstruct, pixel by pixel, the two missing colored information, via a “demosaicing” algorithm. An example of such an algorithm is the paper by Alleyson et al entitled "Linear de-mosaicing inspired by the human visual system" published in April 2005 in IEEE Transactions on Image Processing 14 (4), 439-449. The signal processing module and the processing module 16 may be integrated into a single module or may be separate modules. The acquisition device 1 advantageously makes it possible to obtain color images Iy;s having a good resolution of the green component which is important in particular for algorithms operating on skin images, such as face recognition algorithms. The acquisition device 1 makes it possible at the same time to obtain infrared images I;r. The acquisition system 100 here comprises two artificial light sources 20, 30 illuminating said scene, the first artificial light source 20 emitting a light signal in the infrared and the second artificial light source 30 emitting a light signal in the visible. Said artificial sources 20, 30 are positioned so as to illuminate the scene, and in particular the object 150, here a body part, represented in the form of a face. The first artificial source 20 of infrared light is dimensioned to allow a signal to be provided on the sensor 14 that can be equivalent to the signal generated by very strong visible light, for example approximately 1000 lux. The infrared lighting can therefore be controlled so as to be modulated in intensity between zero and the equivalent of approximately 1000 lux. This first artificial source 20 of infrared light consists, for example, of one or more 850 nm LED light-emitting diodes, the spectrum of which extends approximately from 800 to 875 nm. The second artificial source 30 of visible light, also called a white lighting source, is sized to provide a sufficient signal in complete darkness without generating glare for the user; it generates illumination of, for example, 50 to 100 lux at a usage distance of approximately 70 cm. [Fig. 2] schematically illustrates an example of hardware architecture of the processing module 16. The processing module 16 then comprises, connected by a communication bus 160; a processor or CPU (“Central Processing Unit” in English) 161; a random access memory; RAM (“Random Access Memory” in English) 162; a ROM (“Read Only Memory” in English) 163; at least one communication interface 165 allowing for example the processing module 16 to communicate with the sensor 14 of the acquisition device 1. Optionally, the processing module 16 comprises a storage unit 164 such as a hard disk or a storage medium reader, such as an SD (“Secure Digital” in English) card reader. The processor 161 is capable of executing instructions loaded into the RAM 162 from the ROM 163, from an external memory (not shown), from a storage medium (such as an SD card), or from a communications network. When the processing module 16 is powered on, the processor 161 is capable of reading instructions from the RAM 162 and executing them. These instructions form a computer program causing the processor 161 to implement the methods described in relation to the following figures. Preferably, module 16 has a centralized architecture, but it can also be implemented in the form of a distributed architecture, both in terms of data or components, hardware or software, particularly in the case of dematerialized IT resources. The methods described in relation to the following figures can be implemented in software form by executing a set of instructions by a programmable machine, for example a DSP (“Digital Signal Processor” in English), a microcontroller or a GPU (graphics processor, “Graphics Processing Unit” in English terminology), or be implemented in hardware form by a machine or a dedicated component, for example an FPGA (“Field-Programmable Gate Array” in English) or an ASIC (“Application-Specific Integrated Circuit” in English). [Fig. 3] illustrates a method for acquiring an infrared image and a color image according to an embodiment applied to an acquisition system 100 comprising in particular a sensor 14 comprising only three photosites as well as an infrared light source 20 and a visible light source 30. It should nevertheless be noted that in another embodiment the system could not comprise a visible light source 30, the infrared light source 20 being in this case the only one controlled according to the method, for example in the case of an acquisition system whose usage conditions exclude complete darkness. The first step represented here of the acquisition method 300 is an initialization step 301 setting the exposure parameters, such as gain and / or exposure time, and intensity parameters of the at least one artificial light source of the acquisition system 100, to predetermined values. Preferably, the initialization step 301 also includes a sub-step for determining the target logic levels (in English "digital level") of the infrared and visible signals, as well as a predetermined target distribution of said signals: CrrlCyisible , However, this sub-step could be part of another step of the method 300, for example the comparison step 306. ©,;sible €t Cr are the target logic levels to which it is desired to control the signal respectively in the visible image and the infrared image. These logic levels are determined according to the use of the images, in particular according to the needs of the algorithms, in particular biometric algorithms (recognition and fraud detection).Knowing that on the photosites of the sensor 14, the visible channels are also sensitive to infrared, the sum of these two instructions must necessarily be lower than the maximum logic level N'nax which can be restored by each photosite, for example N,ra,=1023 in the case of an acquisition device 1 with a 10-bit output. In order not to work with images close to saturation, it is advisable to keep a sufficient margin. For example, these target logic levels are thus determined so as to . that: Cr + C,isibte Ÿ Nmaxl 2. For example Cyisipie = 120 and C7r=240 (for 10-bit images, signal divided over 1024 channels ranging from logic level 0 to 1023), knowing that these values ​​are here dimensioned in relation to low light, that is to say here with a gain of 4, configured here as being the maximum gain so as to limit the noise on the image, because if we have more light these targets will not necessarily be reached but the image quality will be sufficient. This results in a predetermined target distribution of said signals, said target distribution being here Crr / Cyisible the ratio of the infrared logic level to the visible logic level. Once the initialization step is completed, the method 300 for acquiring a color image and an infrared image of a scene takes place on the basis of said exposure parameters. Step 302 of receiving an image acquired from the image acquisition device 1 sensitive to visible and infrared wavelengths. The image is acquired with the parameters defined in the previous iteration, namely: exposure time of the sensor 14, gain of the sensor 14, intensity of the infrared lighting source 20, intensity of the white lighting source 30. To ensure the correct association between the image and these parameters, they are notably stored in the form of metadata associated with the acquired image so that when the image is received, this metadata is included in the transfer of the image. This association can alternatively be carried out using the solution presented in patent application FR2212301. The step of processing said received image itself comprises several steps. The processing separates the infrared image and the visible image and defines at least one signal to be controlled and a related control instruction, said at least one control instruction being an infrared and / or visible signal control instruction according to which at least one artificial source of infrared and / or visible light is controlled. The processing is implemented by the processing module 16 of the acquisition system 100. More specifically, the processing step comprises a step 303 of detecting an object in the image and determining the distance d between the detected object and said acquisition device 1. The detection of objects in the image is for example carried out in a known manner by means of an artificial intelligence algorithm or by a neural network, as described for example in the article “For only look once: Unified, real-time object detection”, by Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016), Proceedings of the IEEE conference on computer vision and pattern recognition (p. 779-788). The object 150 sought is for example here a face, and the choice of the face of interest is for example the closest face. A region of interest ROI is defined, for example rectangular, corresponding to the chosen face. The image used for implementing this step 303 of detection and determination of distance is preferably the raw image I; (composite) or the color image 1,;,, but it could also be the infrared image I;r. The distance d from the face 150 to the device is also estimated, for example as a function of the size of the object on the image in pixels, and knowing the intrinsic characteristics of the camera (focal length and pixel size in particular, which are known parameters, or may also be part of the metadata associated with the image during its acquisition). This estimation of distance d is approximate but sufficient; another more precise method of estimating the distance d will be defined later. Preferably, in the event of absence of object detection in the first received image, a loopback represented by a dotted line is carried out to the initialization step 301 which carries out an iterative scan of configurations with image acquisition and analysis until the detection of the object sought in the acquired image. Each configuration comprises said exposure parameters, such as gain and / or exposure time, and intensity parameters (expressed in the table below in duty cycle, ranging from 0 to 1) of the at least one artificial light source, the values ​​of said parameters are fixed at predetermined values ​​in each configuration. Thus, if the object sought is a face, the goal is to provide images that allow a possible face to be detected as soon as possible, regardless of the ambient lighting conditions, and in particular backlighting for which a simple automatic gain control based on the average of the complete image or a fixed region would not be sufficient. An initialization mode is defined here according to the object sought, so in face mode, the acquisition system is configured with a set of parameters optimized for face detection in a given lighting condition, and at each successive image, the parameters will be modified to adapt to various conditions, until a face is detected on one of the images. All the configurations can be grouped in a table, for example: [Tables1] Set of [Time Gain Intensity in- parameters No. |of infrared exposure 0.000625s |1 0 001 s 1 0 0.02 s 8 01 For example, configuration 1 allows a face to be detected with very strong illumination, in the case of direct sunlight for example; Configuration 2 allows the detection of a face with intermediate illumination, in the case of indirect illumination by the sun or strong ambient artificial lighting (i.e. not emitted by the artificial sources of the system) for example; Configuration 3 allows a face to be detected in low light or in complete darkness. To this end, in this configuration the infrared lighting source 20 is switched on at a level that provides the necessary signal. We chain the acquisition of an image with configuration |, then another with configuration 2, then another with configuration 3, then we start again with configuration 1, until a face is detected. If the object sought is for example a visual code, then the set of configurations will be different, adapted to the visual code mode. Then once the object is detected, the method 300 continues with the processing optimization phase from the detected object 150. In certain cases, there may be temporarily no object in the capture volume, which results in an absence of detection during step 303 and in this case, the system 100 can for example freeze its parameters until the next detection, and for example after a given time threshold (1 second for example) loop back to the initialization step with configuration scanning. The processing step comprises the step 304 of calculating at least two metrics on all or part of the image, the first metric characterizing the infrared signal, the second metric characterizing the visible signal. Said part of the image is in particular the region of interest ROI of the image delimiting said detected object, said object being in particular a body part, such as a face, or information, such as a visual code (eg Quick Response Code, Machine Readable Zone), in particular printed on paper or displayed on an e-reader with “e-ink” technology. Each metric can be at least the average over the part of the image corresponding to the ROI, on each wavelength channel. For example, let's take Syr as the average of the infrared channel, and Svisipte as the average of the visible channel. In the case illustrated here of a 1 RVPIR device with a 14 sensor with only three photosites, the visible signal has two components, red and green, and we can take, for example: Syisible = M8X(Syouges Syer1)- In the case of a 1 RGBPIR device with a 14 sensor with four photosites, we would take, for example: S ip, = MAX(Syorge Sverr Spteu) In the case of a magnification objective, once the detection of the object 150 in the first image has been carried out, a second image can be acquired with a magnification so that the entirety of the second image corresponds to the ROI of the first image, this implementation is nevertheless delicate in the case of a person in movement because the displacement of the object must also be taken into account. A step 305 is then carried out for estimating the logic level of the signal due to the ambient infrared. To do this, a step is executed for estimating a component S;rambient of the first metric Syp characterizing a quantity of infrared signal due to the ambient infrared, that is to say not emitted by the sources 20, 30 of the system 100. Indeed, the first metric S;p characterizing the infrared signal is based on an estimation of the logic level of the signal due to the infrared present and knowing that the infrared signal on the image is the sum of the signal due to the ambient infrared illumination external to the product, and of the signal due to the infrared illumination emitted by the system, we obtain: - Calculation of the additional signal S; Additional produced by the infrared lighting source 20 S rh — FrDCrTrmGain with d the distance to the object 150 previously estimated in step 303, expo and Gain: the exposure time (in seconds) and the gain (unitless) applied during the acquisition of the image, and R; z the system parameter characterizing a sensitivity of the optical chain (m? / s) and DC; the current duty cycle of the infrared lighting source 20 (unitless, between 0 and 1); - Calculation of the ambient infrared signal Syrambiant That is to say not emitted by at least one artificial source 20,30 of infrared and / or visible light (knowing that the source 30 of visible light does not emit in the infrared): Srrambiant 7 SIR- SrRadditionnel - Then a calculation step (not shown) of ambient lighting distribution takes place, in particular in the form of the ratio Stganpan of the quantity of the component of the first metric Sisibie characterizing the infrared signal due to the ambient infrared on the second metric. We then proceed to step 306 of comparing said ambient lighting distribution Sirsnbim with said predetermined target distribution _Cr_ during step 301 Seauble Casible initialization, and the control instruction of the infrared signal is a function of said comparison, and in particular in that the control instruction of the visible signal is a function of said comparison. In fact, depending on the said comparison, we decide whether we are going to try to control the visible signal from source 30 or rather the infrared signal from source 20: - If SR aubion «< £R. then we are in the nominal case where there is less infrared Svisble Cossible than visible. In this case, we will control the visible signal to the predetermined visible target logic level, becoming the visible instruction; - Otherwise, there is more infrared than visible, we then control the infrared signal on the infrared setpoint. In the following, we define S the signal that we are going to control, which is equal to SyRembiant OR Svisible according to the criterion thus defined, and C the setpoint to which we are going to control it, that is to say respectively the logic levels Cyr or Cçisibie according to the same criterion. Sometimes a setpoint can already be reached, particularly intrinsically depending on the proximity of the place to a window for example. Finally, the processing step includes: - a step 307 of calculating the exposure parameters of the acquisition device 1 to be applied, such as gain and / or exposure time, in particular as a function of the ratio of the at least one servo-control instruction to said related signal to be servo-controlled, to reach said at least one servo-control instruction; - a step 308 of calculating the control command for the artificial infrared light source; - a step 309 of calculating the control for piloting the artificial source of white light, i.e. visible light. More specifically, in step 307 of calculating the new exposure parameters: We possibly correct S by a non-linear law (typically a polynomial, for example of degree 3) allowing to take into account and compensate for possible signal saturation, this corrected value is called S'. We calculate the ratio Ratio=C / S" which corresponds to the gain that must be applied to the image analyzed to reach the target; We restrict Ratio between an upper bound ratio_min, set for example to 0.1 and an upper bound ratio_max, set for example to 10; We deduce the new product E = T,.po-Gain.Ratio. It is then a question of distributing Texpo' and Gain' in such a way as to have Texpo Gain' = E, in following a strategy of optimizing the gain or the exposure time, which is de- pending the desired performance (in terms of motion blur and noise in particular). Indeed, exposure time and gain act in a proportional on the image, resulting in synergy and their determination then relies on a compromise, because the lengthening of the exposure time can create a crazy move (a maximum crazy move that can be defined in function of the application and the associated algorithm) while the increase gain causes increased noise. For example, we first determine the Tepo exposure time as well. as large as possible while remaining less than E, within the limit of a time maximum exposure (for example 20ms or 40 ms), and we deduce then Gain'=E / Têxpo 3 We can also advantageously favor times exposure times which are multiples of the lighting period ambient (e.g. 10 ms in Europe and 8.33 ms in the USA) of America) to avoid flickering effects due to non-continuous lighting; You can also avoid updating these configuration settings of the sensor 14 in the case where the variation with respect to the image previous would be very low (e.g. variation <10%), which allows for a more stable image in terms of brightness. In fact, Since measurements are noisy, small variations may be due to a measurement noise (e.g. on distance estimation) more than a real variation in ambient brightness. More precisely, during step 308 of calculating the control command for the artificial infrared light source 20, the desired additional infrared signal S is first calculated for the next image from the target infrared setpoint Cr (infrared target logic level), and the ambient infrared signal S, cam that is anticipated for the next image, knowing the current ambient infrared signal S7R ambient €t the ratio Ratio that will be applied to the exposure of the device 1: Ambient Srr = Ratio.S7r ambient * C2 VL gives Additional STr = Crp-RatioS;r embiant Then we calculate the new duty cycle DC,p For infrared lighting according to: * _ Being? DC,p = 55e During step 309 of calculating the control for driving the artificial white light source 30, given the desired level of quality in terms of noise on the image for the exploitation of visible images, and knowing that the noise increases with the gain of the device 1 (at a given output signal level), it is necessary not to exceed a certain gain of the device 1. Here, the control of the artificial white light source 30 is binary, all or nothing, i.e. on or off, for reasons of simplification, but its intensity could alternatively be modulated. Thus, - if the white lighting source 30 was switched off and Texpo Gain' > Threshold,numage Then we switch on the white lighting source 30, with for example the switching-on threshold Switch-on threshold = À x 0.02 (gain of 4 and exposure time of 20 ms); - if the white light source 30 was on and Texpo Gain' <seuilexinction alors the white lighting is switched off, the extinction threshold Seuilexinction Being chosen so that switching on the white lighting on the observed object does not cause a drop in gain sufficient to move to the next iteration, i.e. the next image acquisition, below the extinction threshold, which would cause flashing and instability of the system, for example Seuileincion = 1 X 0.02 (gain of 1 and exposure time of 20 ms), with lighting generating approximately 50 lux on the object. From the moment the lighting is switched on, it is also advantageous not to switch it off until a certain period of time has elapsed, in order to avoid discomfort for the user. Preferably, the two control commands for the artificial infrared light source 20 and the artificial white light source 30 are defined at each iteration. In the above description of the acquisition method 300, only the value of the ambient infrared signal, i.e. not emitted by the artificial source 20 of infrared light nor by the artificial source of visible light 30 of the system 100, is calculated because the modulation of the infrared signal is favored over the modulation of the visible signal so as to avoid switching on the visible lighting source 30. However, one could also calculate the value of the ambient visible signal S visible ambient NON Emitted by the sources 20,30 of the system 100 (in particular 30 here because the source 20 intrinsically only generates infrared) and calculate the additional signal S visibleadds|PTO produced by the lighting source 30 so as to avoid the deviation when the visible light source 30 is switched off and the related reconvergence time. This 300 acquisition process by iterative control allows: to optimize the exposure time and sensor gain, depending on visible ambient illumination to reach a target level Cyisipe SUR l'image visible b. continuously adjust the intensity of the infrared illumination so that the image in- infrared reaches a specific target C,p (in addition to the component infrared already present in ambient lighting) € control the ignition of the visible light source 30 (here in all or nothing) in the case where it detects a lack of visible light, and controls its extinction when visible lighting becomes sufficient again. [Fig. 4] schematically illustrates an image acquisition system according to another embodiment of the invention. Indeed, the acquisition system 100 here comprises not only the acquisition device 1 from which a visible image I and an infrared image Ir originate, but also comprises another visible image acquisition device 2, this other device is only sensitive to the visible, such as an RGB camera, operating at the same time as the image acquisition device 1. RVPIR. The exposure parameters of this other device 2: exposure time and gain are then controlled according to an algorithm independent of the RVPIR image acquisition device 1. The image acquired by this other device 2 is notably used to provide video feedback to the user and for biometric recognition. Furthermore, by coupling this image with the visible image of the RVPIR image acquisition device 1, this makes it possible to detect planar frauds by stereoscopy. A 3D model of the targeted object is obtained. This 3D model is used to detect planar frauds, or those whose overall shape does not correctly approximate that of a face. The spacing e between the two devices 1, 2 is approximately 16 cm, which makes it possible to detect planar frauds up to approximately 90 cm. It is then of interest to position the infrared lighting source 20 with an angle a greater than 5° relative to the RVPIR image acquisition device 1, seen from the face 150. The interest is to not generate a "red eye" effect (in English "bright pupil") which would be dependent on the distance of the person, and would introduce variability into the images, and therefore reduce the performance of the fraud detection. Given here the depth of the capture volumes (distance d of the order of approximately 40-90 cm) this implies a distance D of 6.5 cm to 7.5 cm between the RVPIR image acquisition device 1 and the infrared lighting source 20. Advantageously, the infrared IR channel, or all the channels of the second camera, are processed by artificial intelligence, in order to detect fraud such as 3D masks. The combination of these treatments makes it possible to detect fraud of all types and ensure biometric recognition in all lighting conditions. Stereoscopy also allows the distance d to be determined precisely. The processing module 16 then communicates in particular via its communication interface 165 with the cameras of the two devices 1, 2. [Fig. 5] illustrates an authentication device A according to one embodiment. The authentication device A of a body part comprises an image acquisition system such as the acquisition system 1 described with reference to FIG. 1 and an image analysis module 3. The image processing module 3 comprises a module 30 for recognizing a body part, such as a face, and a module 32 for detecting fraud. It may optionally comprise a module 34 for activating access to a location, e.g., a building, a room, etc. The body part recognition module 30 is connected to a database 4. The database 4 may be part of the authentication device A or be external to it. The database stores images or image descriptors of body parts of authorized persons, e.g. persons authorized to access a building. [Fig.6] illustrates an authentication method as implemented by the authentication device A of [Fig.4]. The user wishing to be authenticated presents his face in front of the image acquisition system 1. During a step S30, a raw image of a face 150 illuminated by infrared and visible light is obtained by the acquisition system 1. In a step S32, a color image L,;, and an infrared image Ir are obtained from the optimized configuration parameters of the acquisition system 1. Known demosaicing algorithms can be applied to obtain the missing values ​​of each pixel. Such an algorithm makes it possible to obtain for each pixel of the color image a value R and a value V and for each pixel of the infrared image an IR value. In a step S34, a face recognition algorithm is applied. For example, the color image I,; obtained is processed, e.g. segmented, to extract a face. The extracted face is then compared with the face images or their descriptors stored in the database 4. If the extracted face is close in the sense of a certain metric to a face in the database 4 then the extracted face is recognized, otherwise the face is unknown. In a variant, the face recognition algorithm uses the color image L, and the infrared image lp. In the case S36 where the face is unknown, access is refused in step S38. In a step S40, a fraud detection algorithm is applied. For this purpose, the infrared image I is used. In a variant, the infrared image I, and the color image I, are used. A neural network can be used for this purpose. In case S42 where fraud is detected, i.e. the face is not authentic (e.g. use of a mask), access is refused at step S38. Access is validated during a step S44 only if the face 150 is recognized and authenticated. In a particular embodiment, an image j' is generated by concatenating the images Ii and I. For this purpose, a demosaicing process is applied to the images Ir and Li, Le. after correction by the calibration parameters. The image la therefore has 3 components R, V and IR. This image la is then used during steps S34 and S40 by face recognition and fraud detection algorithms. These algorithms can use convolutional neural networks having learned their coefficients on images of the type of image l'. For this purpose, resnet type networks can be used. Such networks are notably described in the document by He et al. entitled " Deep residual learning for image recognition” and published in 2016 in Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778). In another embodiment, the color image L,;, with the two components R and G is colorized in order to recreate the blue component and thus obtain an image I" having 3 components R, G and B. The method described in the document by Zhang et al. entitled "Colorful image colorization" published in ECCV in October 2016 is an example of a colorization method based on convolutional neural networks. This solution allows the application of traditional face recognition algorithms that typically use RGB color images. The algorithm described in the paper by Parkhi et al. entitled “Deep face recognition” and published in September 2015 in BMVC (Vol. 1, No. 3, p. 6) is an example of such a face recognition algorithm. The image I” is used by the face recognition algorithm at S34. The fraud detection algorithm can use at S40 either only the image Ir, or both the image [x and the image I”. For this purpose, the algorithm described in the paper by Agarwal et al. entitled “Face presentation attack with latex masks in multispectral videos” and published in 2017 in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops (pp. 81-89) can be used. Colorization is particularly useful for viewing or exporting images. Preferably, the entire authentication process is implemented by the processing module 16, knowing that this module may or may not be centralized.< / seuilexinction>

Claims

Claims

1. System (100) for acquiring a color image (L;s) and an image (In ) infrared of a scene, said acquisition system comprising: - an image acquisition device (1), said device being sensitive at visible and infrared wavelengths, including: - a sensor (14); - a lens (10) configured to focus the light received at the input of the acquisition device (1) on the sensor (14), a double-pass filter (12) strip being placed in front of the sensor (14), and in particular between the lens (10) and said sensor (14); - a processing module (16) connected to said acquisition device (1) image configured to separate, at the sensor output (14), the in- infrared (I,x) and the visible image (I,;,) and to define at least one signal to control and a related control instruction (C), said at least a control instruction being a control instruction of infrared signal (C;z) and / or visible (Cyisibie) €N function of which is controlled at least one artificial light source (20,30); - at least one artificial source (20, 30) of light illuminating said scene, said at least one source comprising an artificial source(20) infrared light emitting a light signal in the infrared and / or an artificial source (30) of visible light emitting a signal luminous in the visible.

2. Acquisition system (100) according to the preceding claim ca- characterized in that said image acquisition device (1) comprises a filter (12) dual bandwidth including: - in the case of a sensor (14) with at least four different photosites, the first bandwidth of the filter (12) is configured to pass visible wavelengths, particularly below 650 nm, and in sensor case (14) with three different photosites, the first band pass-through of the filter (12) is configured to attenuate the blue component of light, notably with a first band bandwidth extending for example between 530 to 650 nm; and - the second bandwidth is configured to allow the infrared wavelengths, especially the second bandwidth extends from 800 to 875 nm.

3. Acquisition system (100) according to the preceding claim ca- characterized in that the second bandwidth extends from 760 nm-800 nm.

4. Acquisition system (100) according to any one of the claims previous characterized in that: -the artificial source (20) of light emitting in the infrared is configured to transmit, within a range of operating distance, in particular from 40 to 100 em, a signal equivalent to the signal generated by intense ambient outdoor visible light, notably 1000 lux, by example a light-emitting diode at 850 nm, whose spectrum extends from 800 to 875nm and / or -the artificial source (30) of light emitting in the visible is configured to illuminate satisfactorily in complete darkness, in particular 50 lux at a usage distance of approximately 70 cm.

5. Acquisition system (100) according to any one of the claims previous characterized in that it comprises another device (2) of shooting only sensitive in the visible, synchronized on the acquisition device (1) and whose exposure parameters are controlled independently of those of the acquisition device (1).

6. Method (300) for acquiring a color image (I,;,) and an in- infrared (Ir) of a scene, said acquisition method comprising steps of: - receiving (302) an image (1,) from a device (1) image acquisition sensitive to visible and in- wavelengths infrared; ; - processing of said received image (Is) separating the infrared image and the visible image and defining at least one signal to be controlled and a related control instruction, said at least one instruction (C) servo being a signal servo instruction in- infrared (Cyr) and / or visible (Cyisibie) €N function of which is controlled at at least one artificial source (20, 30) of infrared and / or visible light.

7. Acquisition method (300) according to claim 6 characterized in that that the processing step comprises a step (303) of detecting a object (150) in the eV / or infrared color image and determination of distance (d) between the detected object (150) and said acquisition device {L).

8. Acquisition method (300) according to any one of claims 6 to 7 characterized in that the processing step comprises a step (304) calculation of at least two metrics (Sir, Svisibie) on all or part of the image, the first metric (S,r) characterizing the infrared signal, the second metric (S,;sibie) Characterizing the visible signal.

9. Acquisition method (300) according to claims 7 and 8 characterized in that said part of the image is a region of interest of the image de- limiting said detected object, said object being in particular a part bodily (150) or information, such as a visual code, in particular printed on a physical document.

10. Acquisition method (300) according to any one of claims 8 to 9 characterized in that the at least two metrics (Sir, Svisibie) SONt calculated as being at least the average over all or said part of the image on each wavelength channel of said characterized signal.

11. Acquisition method (300) according to any one of claims 8 to 10, characterized in that the treatment step comprises: - a step (305) of estimating a component (Sirambiant) of the first metric characterizing a quantity of infrared signal due to ambient infrared, i.e. not emitted by at least one ar- (20,30) of infrared and / or visible light, and; - a step of calculating ambient lighting distribution (Six simple} Suitable especially in the form of the quantity ratio of the component of the first metric characterizing the infrared signal due to infrared ambient from non-artificial source on the second metric, and; - a comparison step (306) of said distribution (Sirantan \ to a Sy sible predetermined target distribution (_Em_), Chusible and in that the infrared signal control setpoint (Cr) is function of said comparison, and in particular in that the instruction ( Crisbh je) of servo control of the visible signal is a function of said com- Parison.

12. Acquisition method (300) according to any one of claims 6 to 11 characterized in that the treatment step comprises: - a step (307) of calculating the exposure parameter of the device (1) acquisition to be applied, such as gain and / or exposure time, in particular depending on the ratio of at least one instruction ( Cr CJ of control on said signal to be controlled relating, for 9 7th you achieve said at least one control instruction; and / or - a calculation step (308, 309) for controlling the source piloting (20,30) artificial infrared and / or visible light.

13. Acquisition method (300) according to any one of claims 6 to 12 characterized in that it comprises an initialization step (301) ef- performing an iterative scan of configurations until the detection of the object, each configuration including at least one parameter exposure, such as gain and / or exposure time, and a parameter intensity of at least one artificial light source, the values said parameters being fixed at predetermined values ​​in each configuration.

14. A method of authenticating a body part characterized in that that it includes: - acquire a color image (1,.) and an infrared image (Ir) of said body part by the acquisition method (300) according to one of the re- claims 7 to 13 and in that the detected object is a body part {150); and - authenticate the body part from said color images (L,;s) and infrared (IR).

15. A device (A) for authenticating a body part, comprising: - an image acquisition system (100) according to one of claims 1 tosS5:and - the processing module (16) being configured to detect a part body (150) from the received image and to perform an authentication fication of said body part detected from said images color (L,;;) and infrared (Ir).