Colour-image acquisition device and associated colour night-vision system

Abstract

The invention relates to an image acquisition device comprising an image intensifier tube; and a CMOS sensor. The CMOS sensor has a sensitive area configured to transform photons transmitted by an optical output interface of the image intensifier tube into images. Furthermore, the CMOS sensor comprises a second set of pixels, each pixel of the second set of pixels of the CMOS sensor being configured to be assigned to one or more pixels of a first set of pixels of a wavelength-selective filter, fastened to an outer face of an input window of the image intensifier tube, by way of a specific calibration phase (Pc).

Description

[0001] COLOR IMAGE ACQUISITION DEVICE AND SYSTEM

[0002] COLOR NIGHT VISION ASSOCIATED

[0003] FIELD OF THE INVENTION

[0004] The invention relates to the field of vision in a very dimly lit environment. Thus, the invention relates to a device conventionally called a “color night vision device”.

[0005] The invention also relates to a specific element of a color night vision device: the image intensifier tube. This is intended to enable, in a known manner, the luminosity of a scene to be amplified. More particularly, the invention aims to observe a night scene in color.

[0006] The invention finds application in many fields where it is sought to improve vision in dark environments.

[0007] In the military and defense field, the invention can be implemented for surveillance applications, detection, recognition and identification of targets, navigation, piloting in low-light conditions. The invention can be integrated into optical devices such as night vision binoculars and night weapon sights. In the aeronautics sector, the invention is applicable to assist pilots during landing or takeoff phases in low-light conditions and for night surveillance operations. In astronomy, the invention can improve the observation capabilities of optical instruments during observations in low-light conditions.

[0008] In search and rescue operations, the invention is beneficial for searching for lost people or animals in nighttime conditions and for navigating in dimly lit areas. In the fields of biology and ecology, the invention allows for the observation of wildlife in nighttime conditions without disturbance from artificial light. In industrial and medical fields, the invention finds application in certain endoscopy procedures where lighting is limited. For hunting and outdoor activities, the invention aids in the detection of game in low light and in navigation during nighttime expeditions. In scuba diving, the invention can be used to explore deep and dark ocean areas. In the fields of photography and cinematography, the invention can allow for the capture of images without the need for intense artificial lighting.

[0009] This invention therefore presents a multitude of potential applications.

[0010] STATE OF THE ART

[0011] As schematically illustrated in Figures 1 and 2a of the state of the art, a night vision device 110 is conventionally in the form of a goggle intended to be placed in front of a sensor, the eye of a user, or both eyes when two devices are juxtaposed to form monocular or binocular night vision binoculars.

[0012] This night vision device 110 integrates several elements placed on the optical axis of the sensor or the user's eye in order to transform the image of the observed scene. More precisely, the scope comprises, from the scene outside the sensor or the user's eye, an objective 12, an image intensifier tube 130 and an eyepiece 14.

[0013] The objective 12 conventionally comprises one or more lenses for capturing the photons of the electromagnetic radiation of the observed scene. The eyepiece 14, similarly to the objective 12, comprises one or more lenses for capturing and incidentally viewing the photons of the light signal emitted by the image intensifier tube 130.

[0014] The image intensifier tube 130 comprises at least three distinct elements: a photocathode 16, an electron multiplier 18, and a phosphorescent screen 20.

[0015] The photocathode 16 is in the form of a semi-transparent photosensitive layer receiving the photons of the incident electromagnetic radiation, that is to say the photons coming from the observed scene and transmitted by the objective 12. To do this, an entrance window 15 transmits the photons from the objective 12 onto the photocathode 16 while ensuring the hermeticity of the external wall 23 of the image intensifier tube 130. The photocathode 16 is generally produced in the form of a thin layer of metal or semiconductor affixed to a layer of glass or material transparent to light. The material of the photocathode 16 is chosen according to its sensitivity to the image of the observed scene. The interaction of the photons of the incident electromagnetic radiation of the observed scene with the photocathode 16 produces, by photoelectric effect, an emission of electrons.

[0016] The emitted electrons, called primary electrons or photoelectrons 28, are then subjected to a first electric field within a first acceleration zone 17 making it possible to direct and accelerate the photoelectrons 28 towards the electron multiplier 18. This first electric field is produced by applying a voltage between the photocathode 16 and the electron multiplier 18, typically a voltage of the order of 50 to 500 volts to guarantee the most rectilinear path possible for the photoelectrons 28.

[0017] The electron multiplier 18, also called an electron amplifier, typically comprises a microchannel plate 25 covered by electrodes. This microchannel plate is also known by the acronym GMC or MCP for "microchannel plate" in the English literature. It is made from a plate of resistive or dielectric material with a thickness typically between 0.2 and 1 millimeter.

[0018] As illustrated more precisely in Figure 2b of the prior art, the microchannels 25 pass right through the wafer of the electron multiplier 18 and are parallel to each other. They have a diameter of between 3 and 12 micrometers. Their internal wall is conventionally coated with a thin semiconductor layer produced by chemical treatment and allowing the transmission of an electric current by jumps. These microchannels 25 have an axis of revolution «2 inclined at an angle «3 of a few degrees, typically between 4 and 12 degrees, relative to the normal of the surface of the electron multiplier 18, so as to induce multiple collisions of the photoelectrons 28 in the microchannels 25.

[0019] In addition to the first electric field created between the photocathode 16 and the electron multiplier 18, a second electric field is created between the two faces of the electron multiplier 18 by means of the electrodes placed on either side of the microchannel plate 25. This electric field makes it possible to charge the internal semiconductor layer of the microchannels 25 so that the multiple collisions of the photoelectrons 28 in the microchannels 25 generate a large number of secondary electrons 29.This electric field also makes it possible, on the one hand, to accelerate the first secondary electrons 29 inside the microchannels 25, so that these first secondary electrons 29 collide again with the surface of the microchannels 25, in turn producing other secondary electrons 29, and so on, generating by this physical process a large number of secondary electrons 29; and on the other hand to accelerate the secondary electrons 29 by supplying energy, in order to orient them from the entrance of the microchannels 25 to the exit of the microchannels 25. Typically, the photoelectrons 28 are multiplied by a factor of between 10. 2 and 10 4 in the electron multiplier 18.

[0020] At the exit of the microchannels 25, these secondary electrons 29 are then moved linearly towards the phosphorescent screen 20, within a second acceleration zone 19, under the effect of a third electric field generated between the exit of the electron multiplier 18 and the phosphorescent screen 20, typically an electric field generated by a voltage of between 4 and 10 kV.

[0021] The phosphorescent screen 20 makes it possible to transform the secondary electrons 29 into photons producing a luminous intensity.

[0022] It comes in the form of a phosphorescent layer or a layer of a luminophore material deposited on a substrate, typically glass in the form of a network of optical fibers or a glass block.

[0023] When it is desired to form binoculars or a night vision scope, the image formed at the output of the phosphorescent screen 20 is transmitted to the eyepiece 14 by optical means, conventionally a network of optical fibers 21 making it possible to return the image formed on the phosphorescent screen 20 to obtain a correct visualization of the observed scene.

[0024] To generate the three electric fields, a power supply unit 22 is conventionally arranged around an internal vacuum enclosure 24 constituting the core of the light intensifier tube 130. The night vision system 110 thus formed therefore has optical elements, a power supply unit 22, and possibly a system for viewing the scene observed on the eyepiece 14. When the night vision system 110 is intended for the human eye, it is conventionally in the form of a telescope and the image intensifier tube 130 is cylindrical. In this embodiment, the photocathode 16, the electron multiplier 18, and the phosphorescent screen 20 conventionally have a diameter of between 15 and 30 millimeters.In order to limit collisions between electrons and gas molecules, particularly air, the two acceleration zones 17 and 19 are conventionally placed under vacuum, at a pressure in the ultra-high vacuum range, of the order of 10'. 7 at 10' 10 mbar. To do this, the photocathode 16, the electron multiplier 18, and the phosphorescent screen 20 are encapsulated in the internal vacuum enclosure 24.

[0025] There are also solutions in current night vision systems that offer to visualize the observed scene in color, in particular by means of color filters placed at the input and output of the image intensifier tube. To do this, a correspondence is made between the pixels of the input filter and the output filter, which are perfectly aligned, in order to preserve the color information generated at the input of the color image intensifier tube.

[0026] In these systems, certain physical phenomena, which can be neglected for a classic night vision system, must nevertheless be taken into account when it is desired to visualize the scene in color.

[0027] The first phenomenon to be taken into account is that of optical aberrations inherent in any optical system. These aberrations can appear in several forms: spherical aberration, coma, astigmatism, field curvature, distortion or chromatic aberrations. These optical aberrations generate a certain deviation of the light ray when passing through the entrance window 15.

[0028] At the output of the photocathode 16, the photoelectrons 28 are emitted with an angular distribution and an energy distribution. The electric field applied between the photocathode 16 and the electron multiplier 18 makes it possible to extract and accelerate the photoelectrons 28 perpendicular to the surface of the photocathode 16. However, there always remains a spatial dispersion between the normal to the surface of the photocathode 16 and the emission direction of the photoelectrons 28. In the same way, at the output of the electron multiplier 18, the secondary electrons 29 are emitted with an angular distribution also forming a spatial dispersion. There therefore exists in any image intensification device 130, between the electron multiplier 18 and the phosphorescent screen 20 as well as between the photocathode 16 and the electron multiplier 18, a spatial dispersion of the path of the photoelectrons 28 and the secondary electrons 29 relative to the optical axis al.The deviation of photons due to optical aberrations inherent in the input window 15 is characterized by the measurement of the LSF, for "Line Spread Function" in the Anglo-Saxon literature. On the other hand, the electronic dispersion is characterized by the measurement of the FWHM, for "Full Width at Half Maximum" in the Anglo-Saxon literature; the FWHM representing the width at half-height of the image of a very fine slit, proportional to its spatial widening.

[0029] More precisely, the photoelectrons 28 are distributed in the plane of the input face of the electron multiplier 18 as a function of the spatial dispersion experienced between the photocathode 16 and the electron multiplier 18. In the same way, the secondary electrons 29 are distributed in the plane of the input face of the phosphorescent screen 20 as a function of the spatial dispersion experienced between the electron multiplier 18 and the phosphorescent screen 20.The combination of these two distributions, namely the distribution of the photoelectrons 28 on the plane of the input face of the electron multiplier 18, and the distribution of the secondary electrons 29 on the plane of the input face of the phosphorescent screen 20, are added to obtain a total spatial distribution of the secondary electrons 29 in the plane of the phosphorescent screen 20 which superimposes the color information generated by the different pixels of the input color filter, which are no longer in correspondence with the pixels of the output color filter, resulting in the formation of a blurred and / or monochrome image.

[0030] In order to define the capacity of an image intensifier tube to transmit color information despite the phenomenon of color overlap, reference is commonly made to the measurement of the color ratio, defined by the ratio between the percentage of color actually perceived by the user at the output of the color image intensifier tube and the totality of the color information generated at the input of the tube.

[0031] The state of the art discloses several devices for color display of images generated by an image intensifier tube. As illustrated in Figure 3, document US 5,233,183 proposes to arrange between the input window 15 and the photocathode 16, a first RGB filter 300, as well as a second RGB filter 310 at the optical output interface 210. The device described in this document does not make it possible to achieve satisfactory color vision, because the phenomenon of spatial dispersion of the electrons between the electron multiplier 18 and the phosphorescent screen 20, as well as the optical aberrations of the input window 15, are not taken into account.Thus, to obtain a color image, it is necessary to degrade the resolution so that the size of each point of the image, that is to say the size of the distinct zones of the filters 300 and 310 is greater than the size of the spatial dispersions of the electrons and the deviations of the photons due to optical aberrations.

[0032] As illustrated in Figure 4, document WO 2004 / 019367 proposes positioning two color filters at the input and output of the image intensifier tube, respectively a first RGB filter 301 in front of the photocathode 16 and a second RGB filter 311 after the phosphorescent screen 210. This document also teaches arranging a microlens array 321 within the objective 120, as well as a second microlens array 322 within the eyepiece lens 140. The device described in this document partially and incorrectly takes into account the phenomena of optical aberration and / or spatial dispersion of the electrons, such that such a vision system also has an unsatisfactory resolution.

[0033] Thus, when it is sought to form binoculars or a night vision scope, various solutions are proposed in the prior art in order to allow current night vision systems to view a scene in color, in particular by means of color filters arranged at the input and output of the image intensifier tube. To do this, a correspondence is made between the pixels of the input filter and the output filter, which are perfectly aligned, in order to preserve the color information generated at the input of the image intensifier tube.

[0034] However, in color night vision systems, the perfect alignment of these two filters can present certain mounting difficulties, and require certain precautions in use. On the other hand, the phenomenon of spatial dispersion of the photoelectrons 28 and the secondary electrons 29 can further complicate the correspondence between the pixels of the input and output color filters, the photon coming from a pixel filtering for a particular wavelength of the first color filter being able to be occulted by a pixel filtering for a different wavelength of the second color filter.

[0035] To limit the influence of this spatial dispersion problem, it is conventional to modify the dimension of each of the pixels of the input and / or output filters, either by oversizing or undersizing the size of the pixels of the filters compared to their initially planned size. This modification thus makes it possible to compensate for the spatial dispersion of the electrons when this is measured and characterized, but drastically limits the resolution of the generated image in the case of oversizing of the pixels, or drastically limits the color information captured in the case of undersizing of the pixels. On the other hand, this modification of the size of the pixels, in particular the undersizing of the pixels, generates a certain noise disturbing the formation of the color images at the output of the image intensifier tube 130.

[0036] It is also known to use a CMOS sensor, at the output of the image intensifier tube 130, in order to improve the processing of information. More precisely, the CMOS sensor makes it possible to acquire the photons coming from the phosphorescent screen 20 by absorbing them on a photosensitive area comprising a set of pixels. Thus, each pixel of this photosensitive area generates an electric current proportional to the number of incident photons received. This electric current is then processed and converted into a digital signal.

[0037] However, the presence of a CMOS sensor associated with two color filters at the input and output of the image intensifier tube 130 does not make it possible to compensate for the loss of color information between the two color filters, due to the aforementioned physical phenomena.

[0038] Solutions have also been proposed in the prior art, for example in documents US2019 / 0098264A1 and WO2016 / 046235A1, to do without the second filter at the optical output interface of the image intensifier tube. These solutions propose to achieve a direct correspondence between the pixels of the first color filter at the input of the image intensifier tube and the pixels of the CMOS sensor by means of a calibration phase. However, in these two documents, no teaching is disclosed as to said calibration phase of the CMOS sensor in order to achieve this correspondence. As when it is sought to form binoculars or a night vision goggle, the use of a CMOS sensor therefore suggests oversizing the pixels of the input filter so as to limit the influence of the spatial dispersion problem.

[0039] In order to be able to obtain high resolutions while retaining the maximum amount of color information, it therefore appears necessary to remedy otherwise the loss of color information due to the physical phenomena responsible for the overlap of colors, and to the processing of the color information from the color filter at the input of the image intensifier tube 130. More particularly, it is necessary to provide a solution by producing a color image acquisition device making it possible to generate high-resolution images while limiting the loss of color information, while ensuring its adequate processing in order to optimize the color vision of the observed scene.

[0040] STATEMENT OF THE INVENTION

[0041] The invention proposes to address this technical problem by placing a single wavelength selective filter fixed on an external face of the input window of the image intensifier tube, and a CMOS sensor configured to transform the light information from the phosphorescent screen, the CMOS sensor being subjected to a particular calibration phase in order to preserve as much as possible the color information generated by the single wavelength selective filter.

[0042] With this specific calibration phase, it is possible to increase the number of pixels of the selective filter in order to obtain a better resolution and more detailed color information, thus achieving a better color image quality at the output of the image acquisition device than that generated by existing devices, where this calibration phase is not carried out. Advantageously, the phenomenon of spatial dispersion of photoelectrons and secondary electrons is better taken into account, and better corrected.

[0043] The selective filter may comprise a number of pixels between 20,000 and 8,000,000, preferably between 40,000 and 1,000,000. Preferably, the pitch defined by the distance between two consecutive pixels of the selective filter is between 5 pm and 50 pm, and preferably between 10 pm and 30 pm. Indeed, for a number of pixels less than 20,000, the spectral resolution of the detected color is too small for the use of the device. Above 8,000,000 pixels, and preferably above 1,000,000 pixels, the spatial dispersion through the image intensifier tube limits the spatial resolution in detection.

[0044] The invention stems from the observation that color information can be preserved more efficiently than with the use of two wavelength selective filters, the presence of two wavelength selective filters at the input and output of the intensifier tube cutting off part of the color information received, and degrading the display of color images at the output of a color image acquisition device.

[0045] Indeed, within the framework of the invention, it has been demonstrated that the arrangement of a CMOS sensor at the output of the image intensifier tube comprising two wavelength selective filters at the input and output of the image intensifier tube did not make it possible to improve the quality of the color images generated, since they were deprived of part of the color information.

[0046] On the other hand, it has also been observed that dispensing with a second wavelength selective filter by having a CMOS sensor, and by achieving a correspondence between the pixels of the wavelength selective filter and the pixels of the CMOS sensor, could not allow acceptable conservation of color information without carrying out a particular calibration phase between the wavelength selective filter and the CMOS sensor. Indeed, it has also been demonstrated that a particular calibration phase was essential for the conservation of color information when the intensifier tube only has a single wavelength selective filter and a CMOS sensor.

[0047] Another method was therefore sought to limit the loss of color information and optimize the conservation of this color information, generated at the input of the image intensifier tube thanks to the wavelength selective filter. Thus, instead of conventionally using two color filters at the input and output of the image intensifier tube, it was sought to avoid a partial loss of color information by a second wavelength selective filter due to the color overlap phenomenon.

[0048] With this new approach, it was possible to determine a specific configuration to meet this constraint.

[0049] Thus, according to a first aspect, the invention relates to an image acquisition device comprising:

[0050] - an image intensifier tube comprising the following elements fixed within an internal vacuum enclosure:

[0051] ■ an input window configured to receive and transmit photons;

[0052] ■ a selective filter for at least three distinct wavelengths, comprising a first set of pixels, fixed on an external face of the input window;

[0053] ■ a photocathode capable of converting photons transmitted through the input window into photoelectrons, said photocathode being fixed on an internal face of the input window;

[0054] ■ an electron multiplier, comprising two parallel faces, capable of multiplying photoelectrons into secondary electrons;

[0055] ■ a phosphorescent screen transforming secondary electrons into photons; and

[0056] ■ an optical output interface configured to capture the photons emitted by the phosphorescent screen to transmit them to the output of the image intensifier tube;

[0057] - a CMOS sensor having a sensitive area configured to transform photons transmitted by the optical output interface into an image and comprising a second set of pixels, each pixel of the second set of pixels of the CMOS sensor being configured to be assigned to one or more pixels of the first set of pixels of the wavelength selective filter, by means of a calibration phase.

[0058] According to the invention, said sensitive area of ​​the CMOS sensor has a number of pixels between 8 million and 40 million by means of a calibration phase which comprises the following steps:

[0059] - transmission of radiation of a first wavelength on the input window, the first wavelength corresponding to one of said at least three wavelengths selected by said filter; - determination of a first group of pixels, the least sensitive to said first wavelength, the numerical value of which is less than a first threshold numerical value;

[0060] - transmission of radiation of a second wavelength on the input window, the second wavelength corresponding to one of said at least three wavelengths selected by said filter;

[0061] - determination of a second group of pixels, the least sensitive to said second wavelength, whose numerical value is lower than a second threshold numerical value;

[0062] - transmitting radiation of a third wavelength onto the input window, the third wavelength corresponding to one of said at least three wavelengths selected by said filter;

[0063] - determination of a third group of pixels, the least sensitive to said third wavelength, whose numerical value is lower than a third numerical threshold value;

[0064] - assigning a first group of calibrated pixels of the CMOS sensor to the pixels of the filter selecting said first wavelength, said first group of calibrated pixels corresponding to the pixels common to the second group of pixels and to the third group of pixels;

[0065] - assignment of a second group of calibrated pixels of the CMOS sensor to the pixels of the filter selecting said second wavelength, said second group of calibrated pixels corresponding to the pixels common to the first group of pixels and to the third group of pixels;

[0066] - assigning a third group of calibrated pixels to the pixels of the filter selecting said third wavelength, said third group of calibrated pixels corresponding to the pixels common to the first group of pixels and to the second group of pixels.

[0067] It should be noted that these calibration steps can be implemented by a control unit of the image acquisition device. Thus, the calibration phase advantageously makes it possible to do without a second wavelength-selective filter at the output of the image intensifier tube, by achieving a correspondence between the color information generated by the pixels of the single wavelength-selective filter present at the input of the image intensifier tube, and the pixels of the photosensitive zone of the CMOS sensor by means of a precise and determined calibration phase, this in order to avoid the loss of part of the color information inadvertently filtered by a second selective filter due to the color overlap phenomenon; and in order to optimize as much as possible the conservation of the color information generated by the wavelength-selective filter.

[0068] With this replacement of a second selective filter by a CMOS sensor associated with the calibration phase, the main physical phenomena responsible for color overlap are limited as completely as possible.

[0069] According to one embodiment, in order to limit the loss of light power available at the input of the color image intensifier tube, the wavelength selective filter integrates transparent pixels.

[0070] Preferably, but not exclusively, the selective filter is a SPARSE matrix type color filter, making it possible to maximize the conservation of light power while providing a color image of the observed scene. A SPARSE matrix is, for example, a matrix which comprises at least as many transparent pixels as wavelength-selective pixels.

[0071] According to a preferred variant, the selective filter is a color filter of the Bayer matrix type, RGBW matrix, or any other matrix of colored filters.

[0072] Preferably, in order to improve the processing of color information by the CMOS sensor when the wavelength selective filter comprises transparent pixels, the pixels of the CMOS sensor not assigned to a wavelength following the calibration phase are assigned to the radiation coming from the transparent pixels of the selective filter. Preferably, the image intensifier tube comprises a microlens array fixed on the wavelength selective filter so as to focus, at the output of the input window, the photons in the plane of the photocathode; the focusing of the photons being determined so as to anticipate the optical aberrations of the input window and / or the dispersion of the electrons between the electron multiplier and the phosphorescent screen so as to avoid the overlap of the colors at the optical output interface.

[0073] The distance between two consecutive microlenses defines the pitch of the microlens array. This pitch of the microlens array may be greater than the pitch defined by two consecutive pixels of the selective filter and the CMOS sensor. Thus, the ratio between the pitch defined by the distance between two consecutive pixels of the selective filter and between the pitch of the microlens array may be between 2 inclusive and 8 inclusive, preferably between 3 inclusive and 6 inclusive. The pitch of the microlens array may be between 0.625 pm and 25 pm, preferably between 2 pm and 12 pm and preferentially between 4 pm and 8 pm. Thus, it is possible to limit the relative chromatic aberration (or color ratio) while maximizing the spatial resolution of the device. In particular, the device may exhibit a relative chromatic aberration greater than 0.5 while exhibiting a spatial resolution greater than 50 Ip / mm (even line per millimeter), preferably greater than 801p / mm.

[0074] Preferably, in order to limit color overlap and further improve image resolution, the microlens array has two, three or four microlenses per pixel of the selective filter.

[0075] Preferably, in order to limit color overlap and further improve image resolution, the microlens array has between four microlenses and nine microlenses per pixel of the first wavelength selective filter.

[0076] Preferably, the CMOS sensor is only photosensitive.

[0077] Preferably, in order to obtain a color image in a multitude of spectral bands, the phosphorescent screen is based on P45 type phosphor. According to a second aspect, the invention also relates to a night vision system comprising: a lens, an image acquisition device according to the first aspect of the invention and a screen for displaying the image of said CMOS sensor.

[0078] BRIEF DESCRIPTION OF THE FIGURES

[0079] The invention will be better understood on reading the following description, given solely by way of example, and carried out in relation to the appended drawings, in which identical references designate identical or similar elements, and in which: Figure 1 illustrates a schematic sectional view of a night vision device of the state of the art;

[0080] Figure 2a illustrates a perspective view of the image intensifier tube of the night vision device of Figure 1;

[0081] Figure 2b illustrates a schematic sectional view of the electron multiplier or night vision device of Figure 1;

[0082] Figure 3 illustrates a schematic sectional view of an image intensifier tube in an enclosure of a color night vision device according to a prior art embodiment;

[0083] Figure 4 illustrates a schematic sectional view of an image intensifier tube in an enclosure of a color night vision device according to another embodiment of the prior art;

[0084] Figure 5 illustrates a schematic sectional view of a color image acquisition device, comprising a color image intensifier tube in an enclosure of a color night vision device according to one embodiment of the invention;

[0085] Figure 6a illustrates a schematic view of a set of sixteen square pixels of a selective filter according to one embodiment of the invention;

[0086] Figure 6b illustrates a schematic view of a set of sixty-four pixels of the photosensitive area of ​​the CMOS, receiving two radiations from a filtering pixel for a first wavelength and a filtering pixel for a second wavelength;

[0087] Figure 6c illustrates a schematic view of a set of sixty-four pixels of the photosensitive area of ​​the CMOS, receiving two radiations from a filtering pixel for a first wavelength and a filtering pixel for a third wavelength; Figure 6d illustrates a schematic view of a set of sixty-four pixels of the photosensitive area of ​​the CMOS, receiving two radiations from a filtering pixel for a first wavelength and a filtering pixel for a third wavelength;

[0088] Figure 6e illustrates a diagram representing the steps of the calibration phase corresponding to the transmissions of the radiations of different wavelengths, and to the assignments of the pixels of the CMOS sensor to the pixels of the wavelength-selective filter; Figure 7a illustrates a schematic view of a set of sixteen square pixels of a selective filter according to an embodiment of the invention;

[0089] Figure 7b illustrates a schematic view of a microlens array according to one embodiment of the invention;

[0090] Figure 7c illustrates a schematic view in detail of the superposition of the set of sixteen square pixels illustrated in Figure 7a with the microlens array illustrated in Figure 7b;

[0091] Figure 8a illustrates a schematic view of a set of sixteen square pixels of a selective filter according to one embodiment of the invention;

[0092] Figure 8b illustrates a schematic view of a microlens array according to another embodiment of the invention;

[0093] Figure 8c illustrates a schematic view in detail of the superposition of the set of sixteen square pixels illustrated in Figure 8a with the microlens array illustrated in Figure 8b;

[0094] Figure 9a illustrates a schematic view of a set of sixteen square pixels of a selective filter according to one embodiment of the invention;

[0095] Figure 9b illustrates a schematic view of a microlens array according to another embodiment of the invention;

[0096] Figure 9c illustrates a schematic view in detail of the superposition of the set of sixteen square pixels illustrated in Figure 9a with the microlens array illustrated in Figure 9b;

[0097] Figure 10 illustrates a schematic view of a set of twelve rectangular pixels of a SPARSE type selective filter according to an embodiment of the invention; and

[0098] Figure 11 illustrates a graph of the emission spectrum of the P45 type phosphor constituting the phosphorescent screen. DETAILED DESCRIPTION OF THE INVENTION

[0099] Figure 5 illustrates a color night vision device 11 integrating a color image acquisition device comprising a color image intensifier tube 13 according to the invention. The color night vision device 11 is in the form of an optical system intended to generate color images of a night scene.

[0100] This color night vision device 11 integrates several elements placed on the optical axis al of the optical system in order to transform the image of the observed scene. More precisely, the telescope comprises, from the scene outside the sensor, an objective 12, the image acquisition device 25 comprising the color image intensifier tube 13 of the invention, and a screen 14.

[0101] As described above, the objective 12 comprises one or more lenses for capturing the photons of the electromagnetic radiation of the observed scene. The screen 14 is in the form of a display system for processing and transmitting the images generated by the image acquisition device 25.

[0102] The image acquisition device 25 comprises a color image intensifier tube 13 and a CMOS sensor 26.

[0103] The color image intensifier tube 13 comprises at least three separate elements integrated in an internal vacuum enclosure 24: a photocathode 16, an electron multiplier 18, and a phosphorescent screen 20.

[0104] As described above, the photocathode 16 is in the form of a semi-transparent photosensitive layer receiving the photons of the incident electromagnetic radiation, i.e. the photons transmitted by the objective 12. To do this, an input window 15 transmits the photons from the objective 12 onto the photocathode 16 while ensuring the hermeticity of an external wall 23 of the color image intensifier tube 13. The photocathode 16 is generally produced in the form of a thin layer of metal or semiconductor affixed to a glass layer. The interaction of the photons of the incident electromagnetic radiation of the observed scene with the photocathode 16 produces, by photoelectric effect, an emission of electrons.

[0105] The emitted electrons, called primary electrons or photoelectrons 28, are then subjected to a first electric field within a first acceleration zone 17 making it possible to direct the photoelectrons 28 towards the electron multiplier 18. This first electric field is produced by applying a voltage between the photocathode 16 and the electron multiplier 18, typically a voltage of between 50 and 500 volts to best guarantee a straight path for the photoelectrons 28.

[0106] In addition to the first electric field created between the photocathode 16 and the electron multiplier 18, a second electric field is created between the two faces of the electron multiplier 18 by means of the electrodes placed on either side of the microchannel plate 25. This electric field makes it possible to charge the internal semiconductor layer of the microchannels 25 so that the multiple collisions of the photoelectrons 28 in the microchannels 25 generate a large number of secondary electrons 29.This electric field also makes it possible, on the one hand, to accelerate the first secondary electrons 29 inside the microchannels 25, so that these first secondary electrons 29 collide again with the surface of the microchannels 25, in turn producing other secondary electrons 29, and so on, generating by this physical process a large number of secondary electrons 29; and on the other hand to accelerate the secondary electrons 29 by supplying energy, in order to orient them from the entrance of the microchannels 25 to the exit of the microchannels 25. Typically, the primary electrons 28 are multiplied by a factor of between 10. 2 and 10 4 in the electron multiplier 18.

[0107] At the exit of the microchannels 25, these secondary electrons 29 are then accelerated in a rectilinear manner towards the phosphorescent screen 20, within a second acceleration zone 19, under the effect of a third electric field generated between the electron multiplier 18 and the phosphorescent screen 20, typically an electric field generated by a voltage of between 4 and 10 kV. The phosphorescent screen 20 makes it possible to transform the secondary electrons 29 into photons producing a luminous intensity. It is in the form of a phosphorescent layer or a layer of a luminophore material deposited on a surface of the optical output interface 21, said surface of the optical output interface 21 facing the electron multiplier 18.

[0108] At the output of the phosphorescent screen 20, the image formed is transmitted to the CMOS sensor 26 by an optical output interface 21, conventionally in the form of a network of optical fibers.

[0109] As described above, the CMOS sensor 26 makes it possible to acquire the photons coming from the phosphorescent screen, absorbing them on a photosensitive zone 27 to radiation of a wide range of wavelengths. This photosensitive zone 27 comprises a set of pixels, each pixel of this zone generating a proportional electric current upon contact with an incident photon. This electric current, in the form of an analog signal, is then processed and converted into a digital signal / value. These digital signals / values ​​are finally processed to display the color images generated by the color image intensifier tube 13, which are displayed on the screen 14. The signal processing is done by any appropriate signal and / or image processing means. The screen 14 may be in the form of a camera or a conventional screen.

[0110] The number of pixels in the photosensitive area 27 of the CMOS sensor 26 is between 8 million and 40 million, with a pixel density per square millimeter ranging from 13,000 to 160,000. The spectral sensitivity of the CMOS sensor extends over a wavelength spectrum from 400 to 1000 nanometers.

[0111] To generate the three electric fields, a power supply unit 22 is conventionally arranged around the internal vacuum enclosure 24. The color night vision system 11 thus formed therefore has optical elements, a power supply unit 22, and possibly a system for viewing the scene observed on the eyepiece 14. The color night vision device 11 is conventionally in the form of a telescope and the color image intensifier tube 13 is cylindrical. In the embodiment of Figure 5, the photocathode 16, the electron multiplier 18, and the phosphorescent screen 20 conventionally have a diameter of between 15 and 30 millimeters.The phosphorescent screen is preferably based on P45 type phosphor, having an emission spectrum adapted to detailed color vision in several spectral bands located between 400 and 750 nanometers, and more particularly between 400 and 700 nanometers, as illustrated in figure 11.

[0112] More specifically, still with reference to FIG. 11, the emission spectrum of the P45 type phosphor constituting the phosphorescent screen 20 is shown, emitting a brightness in a plurality of spectral bands between 400 and 750 nanometers. More particularly, this graph represents the intensity of the brightness of the P45 type phosphor as a function of the wavelength, expressed in nanometers. The brightness is expressed as a percentage, and normalized as a function of the lowest brightness and the most intense brightness for each wavelength over the entire spectrum ranging from 400 to 750 nanometers.

[0113] In order to limit collisions between electrons and gas molecules, particularly air, the two acceleration zones 17 and 19 are conventionally placed under vacuum, at a pressure in the ultra-high vacuum range, of the order of 10' 7 at 10' 10 mbar. To do this, the photocathode 16, the electron multiplier 18, and the phosphorescent screen 20 are encapsulated in the internal vacuum enclosure 24.

[0114] As illustrated in Figure 5, a wavelength-selective filter 30, comprising a set of pixels, is fixed to an external face of the input window 15. The optical output interface 21 is integral with the photosensitive zone 27 of a CMOS sensor 26. Preferably, the photosensitive zone 27 of the CMOS sensor 26 is at least larger in surface area than the surface area of ​​the optical output interface 21, such that all the radiation coming from the optical output interface 21 can be received on the photosensitive zone 27 of the CMOS sensor 26.

[0115] The selective filter 30 may comprise a number of pixels between 20,000 and 8,000,000, preferably between 40,000 and 1,000,000. Furthermore, the pitch defined by the distance between two consecutive pixels of the selective filter 30 is between 5 pm and 50 pm, and preferably between 10 pm and 30 pm. The wavelength-selective filter 30 advantageously comprises transparent pixels, said filter 30 being for example in the form of a SPARSE matrix type color filter, that is to say a matrix which comprises at least as many transparent pixels as wavelength-selective pixels. For example, as illustrated in Figure 10, a SPARSE SI type matrix can integrate RGB elementary cells, reproduced as many times as necessary to form the entire filter, with two red pixels R, two green pixels G and two blue pixels B and six transparent pixels T, interposed between the red pixels R, green pixels G and blue pixels B.Alternatively, the filter 30 may also be in the form of a Bayer matrix type color filter, RGBW matrix, or any other matrix of colored filters. Generally speaking, the filter 30 may be in any form of matrix of colored filters according to various matrix compositions which would present a preservation of the sensitivity of the incident radiation thanks to transparent pixels.

[0116] Referring to Figure 6a, the wavelength-selective filter 30 comprises several pixels filtering the incident light for several particular wavelengths. More particularly, the filter 30 is a selective filter for at least three distinct wavelengths k1, k2 and k3. Figure 6a represents by way of illustrative example a subset of sixteen pixels 31 of the filter 30 comprising four pixels selective for a first wavelength k1, four pixels selective for a second wavelength k2, four pixels selective for a third wavelength k3 and four transparent pixels.

[0117] The calibration phase Pc, making it possible to assign the pixels of the photosensitive area 27 of the CMOS sensor 26 to each particular wavelength k1, k2 and k3, comprises several steps, which are illustrated in Figures 6b to 6e. This calibration phase Pc can be implemented by a control unit, not shown, integrating a microcontroller or a microprocessor and a memory for storing the values ​​of the pixels of the CMOS sensor 26 captured at each step.

[0118] The first step 50 consists of illuminating the color image intensifier tube 13 with a first wavelength kl, which is transmitted by the objective 12, the input window 15 and by the pixels of the filter 30 selecting this first wavelength. This radiation, after intensification by the color image intensifier tube 13, is received on the pixels of the photosensitive area 27 of the CMOS sensor 26. The values ​​of each pixel of the photosensitive area 27 of the CMOS sensor 26 are then recovered, in a step 51, and the pixels whose value is lower than a first threshold value SI are searched for. These pixels form a first group of pixels G1 which are the least sensitive to said first wavelength kl.

[0119] In the same way, step 52 aims to illuminate the input window 15 with a second wavelength λ2. After transmission in the color image intensifier tube 13, the values ​​of each pixel of the photosensitive zone 27 of the CMOS sensor 26 are then recovered, in a step 53, and the pixels whose value is lower than a second threshold value S2 are searched for. These pixels form a second group of pixels G2 which are the least sensitive to said second wavelength λ2.

[0120] These illumination and determination steps are again reproduced with a third wavelength 13. Thus, step 54 aims to illuminate the input window 15 with a third wavelength λ3. In step 55, the pixels whose value is lower than a third threshold value S3 are searched for, forming a third group of pixels G3 that are the least sensitive to said third wavelength λ3.

[0121] For the purposes of simplification and understanding of these first steps, figures 6b to 6d illustrate a subset of sixty-four pixels of the photosensitive zone 27 of the CMOS sensor 26, receiving two distinct radiations from two pixels each filtering for a particular wavelength.

[0122] Figure 6b thus represents the pixel groups G2 and G3, determined during steps 53 and 55 of the calibration phase Pc. In the same way, Figure 6c represents the pixel groups G1 and G3, determined during steps 51 and 55 of the calibration phase Pc; and Figure 6d represents the pixel groups G1 and G2, determined during steps 51 and 53 of the calibration phase Pc.

[0123] As illustrated in Figures 6b to 6d, the pixel groups G1, G2 and G3 form intersection zones, in which some pixels of the photosensitive area 27 of the CMOS sensor 26 correspond to two pixel groups among the pixel groups G1, G2 and G3. Figure 6b thus represents an intersection zone in which a first calibrated pixel group G23 belongs to the pixel groups G2 and G3. Similarly, Figure 6c represents an intersection zone in which a calibrated pixel group G13 belongs to the pixel groups G1 and G3. Finally, Figure 6d represents an intersection zone in which a calibrated pixel group G12 belongs to the pixel groups G1 and G2.

[0124] Steps 56, 57 and 58 of the calibration phase Pc correspond to the assignment of the pixels of the photosensitive zone 27 of the CMOS sensor 26 to the pixels of the filter 30 filtering the different wavelengths kl, k2 and k3.

[0125] Thus, the first groups of calibrated pixels G23 are assigned to the pixels filtering for the first wavelength kl, the second groups of calibrated pixels G13 are assigned to the pixels filtering for the second wavelength k2, and the third groups of calibrated pixels G12 are assigned to the pixels filtering for the third wavelength k3.

[0126] With reference to figures 6a to 6d, certain pixels of the photosensitive zone 27 of the CMOS sensor 26 will not have any allocation following the steps of the calibration phase Pc. These pixels are allocated to the radiation coming from the transparent pixels of the filter 30.

[0127] Preferably, and as illustrated in Figure 5 and Figures 7a to 9c, the color image intensifier tube 13 preferably comprises a microlens array 32 fixed to the wavelength-selective filter 30 so as to focus, at the output of the input window 15, the photons in the plane of the photocathode 16.

[0128] To do this, the microlenses 40 are designed with a thickness and a radius of curvature adapted so that the rays incident on these microlenses 40 pass through the input window 15 and have a maximum of convergence on a plane coplanar to that of the photocathode 16. For the purposes of the invention, the focusing in the plane of the photocathode 16 corresponds to a focusing on the plane of an internal surface of the photocathode 16, said surface of the photocathode 16 being opposite the input window 15. The focusing of the photons in the plane of the photocathode 16 is determined so as to anticipate the optical aberrations of the input window 15 and / or the dispersion of the electrons between the electron multiplier 18 and the phosphorescent screen 20, so as to limit the overlap of the colors at the optical output interface 21 without degrading the resolution of the image output from the color image intensifier tube 13.

[0129] The microlens array 32 has a number of microlenses 40 which can be variable. In this case, the microlens array 32 has at least two microlenses 40, three microlenses 40 or four microlenses 40 per pixel of the first wavelength-selective filter 30. Preferably, the microlens array 32 has between four microlenses 40 and nine microlenses 40 per pixel of the first wavelength-selective filter 30.

[0130] Furthermore, the pitch of the microlens array 32 may be between 0.625 pm and 25 pm, preferably between 2 pm and 12 pm and preferentially between 4 pm and 8 pm. Furthermore, the ratio between the pitch defined by the distance between two consecutive pixels of the selective filter 30 and the pitch of the microlens array 32 is between 2 inclusive and 8 inclusive, preferably between 3 inclusive and 6 inclusive.

[0131] The number of 40 microlenses can thus be adapted according to the desired resolution, while taking into account the desired color ratio. The 40 microlenses are ideally square or rectangular in shape.

[0132] The filter 30 may alternatively have square pixels 39 or rectangular pixels 43 or 46. The microlens array 32 is ideally formed so as to have several microlenses 40 per square pixel 39 or rectangular pixel 43 or 46. Ideally, the projected area of ​​the microlenses of each square pixel 39 or rectangular pixel 43 and 46 must represent at least 90% of the total area of ​​said pixel, the projected area being understood as the projected area of ​​the microlens according to the normal of the first filter 30 and / or the second filter 31. According to a first exemplary embodiment illustrated in FIGS. 7a to 7c, the first filter 30 and the second filter 31 have square pixels 39, and the microlens array 32 has four square microlenses 40 per square pixel 39 of the first filter 30 and second filter 31.Thus, in this exemplary embodiment, the length of the side of each square pixel 39 corresponds to at least a multiple of two of the length of the side of the microlenses 40.

[0133] Figure 7b shows a microlens array 32 of the same size as a set of sixteen square pixels 37 from the first filter 30 and / or the second filter 31, illustrated in Figure 7a. Figure 7c thus represents a subset of four square pixels 38 from the superposition of the set of sixteen rectangular pixels 37 and the microlens array 32 illustrated in Figure 7b.

[0134] According to a second exemplary embodiment illustrated in figures 8a to 8c, the first filter 30 and the second filter 31 have rectangular pixels 43, and the microlens array 32 has two square microlenses 40 per rectangular pixel 43 of the first filter 30 and second filter 31.

[0135] Thus, in this exemplary embodiment, the length of each rectangular pixel 43 corresponds to at least a multiple of two of the length of the side of the microlenses 40, while the width of each rectangular pixel 43 corresponds at least to the length of the side of each microlens 40.

[0136] Figure 8b illustrates a microlens array 32 of the same dimension as a set of sixteen rectangular pixels 41 from the first filter 30 and / or the second filter 31, illustrated in Figure 8a. Figure 8c thus represents a subset of four rectangular pixels 42 from the superposition of the set of sixteen rectangular pixels 41 and the microlens array 32 illustrated in Figure 8b.

[0137] According to a third exemplary embodiment illustrated in figures 9a to 9c, the first filter 30 and the second filter 31 have rectangular pixels 46, and the microlens array 32 alternately has three square microlenses 40 per rectangular pixel 46 of the first filter 30 and second filter 31. Thus, in this exemplary embodiment, the length of each rectangular pixel 46 corresponds to at least a multiple by three of the length of the side of the microlenses 40, while the width of each rectangular pixel 46 corresponds at least to the length of the side of each square microlens 40.

[0138] Figure 9b shows a microlens array 32 of the same size as a set of sixteen rectangular pixels 44 from the first filter 30 and / or the second filter 31, illustrated in Figure 9a. Figure 9c thus represents a subset of four rectangular pixels 45 from the superposition of the set of sixteen rectangular pixels 44 and the microlens array 32 illustrated in Figure 9b.

[0139] As illustrated in Figures 7a to 9c, the set of square pixels 37 and the set of rectangular pixels 41 and 44 have at least three different color filters, these filters advantageously being in the form of a blue color filter, a red color filter and a green color filter. These sets of square pixels 37 or rectangular pixels 41 and 44 also have transparent pixels, said transparent pixels also being covered by the microlens array during the assembly of the color image intensifier tube 13 according to the invention.

[0140] With reference to Figures 5 to 9c, the microlens array 32 can be produced by thermal reflow and by photolithography. Other methods of manufacturing the microlens array 32 can be envisaged, such as by hot embossing / embossing, by projection of plastic microdroplets, or by ultra-precision machining technologies. This microlens array 32 can thus be produced directly on the wavelength selective filter 30.

[0141] The filter 30 is conventionally produced in the form of a layer of gelatin deposited on the desired surfaces, in this case respectively on the external face of the entrance window 15 facing the objective 12 and on the face of the phosphorescent screen 20 facing the electron multiplier 18. The layer of gelatin is then divided into several sections by photolithography and colored using various dyes in order to form the different colors of the first filter 30 and second filter 31. Advantageously, the gelatin before coloring has a transmittance of at least 90%, and preferably from 95% to 99% over the entire wavelength spectrum between 400 and 1000 nanometers, in order to allow the transparent pixels maximum transmittance.

[0142] With the joint use of a wavelength selective filter and a CMOS sensor, the pixels of the photosensitive zone of which are configured to be assigned to one or more pixels of the filter according to a particular calibration phase Pc, advantageously supplemented by the presence of a microlens array 32, the invention makes it possible to drastically limit the phenomenon of color overlap, the image acquisition device 25 thus being able to preserve as much as possible the color information generated by the wavelength selective filter, thus generating colored images whose color information is maximum, without degrading the image resolution and without compromising the color ratio, for optimized color night vision.

Claims

CLAIMS 1. Image acquisition device (25) comprising: - an image intensifier tube (13) comprising the following elements fixed within an internal vacuum enclosure (24): ■ an input window (15) configured to receive and transmit photons; ■ a filter (30) selective for at least three distinct wavelengths (I, X2, X3), comprising a first set of pixels fixed on an external face of the input window (15); ■ a photocathode (16) capable of converting photons transmitted by the input window (15) into photoelectrons (28), said photocathode (16) being fixed on an internal face of the input window (15); ■ an electron multiplier (18) comprising two parallel faces (38, 39), capable of multiplying the photoelectrons (28) into secondary electrons (29); ■ a phosphorescent screen (20) transforming the secondary electrons (29) into photons; and ■ an optical output interface (21) configured to capture the photons emitted by the phosphorescent screen (20) to transmit them to the output of the image intensifier tube (13); - a CMOS sensor (26) having a sensitive area configured to transform photons transmitted by the optical output interface (21) into images and comprising a second set of pixels, each pixel of the second set of pixels of the CMOS sensor (26) being configured to be assigned to one or more pixels of the first set of pixels of the wavelength-selective filter (30), by means of a calibration phase (Pc); characterized in that said sensitive area of the CMOS sensor (26) has a number of pixels of between 8 million and 40 million by means of a calibration phase (Pc) which comprises the following steps: - transmission (50) of radiation of a first wavelength (XI) on the input window (15), the first wavelength (XI) corresponding to one of said at least three wavelengths (XI, X2, X3) selected by said filter - determination (51) of a first group of pixels (Gl), the least sensitive to said first wavelength (XI), the numerical value of which is less than a first threshold numerical value (SI); - transmission (52) of radiation of a second wavelength (2) on the input window (15), the second wavelength (2) corresponding to one of said at least three wavelengths (I, X2, X3) selected by said filter (30); - determination (53) of a second group of pixels (G2), the least sensitive to said second wavelength (X2), the digital value of which is less than a second threshold digital value (S2); - transmission (54) of radiation of a third wavelength (X3) on the input window (15), the third wavelength (X3) corresponding to one of said at least three wavelengths (XI, X2, X3) selected by said filter (30); - determination (55) of a third group of pixels (G3), the least sensitive to said third wavelength (X3), the digital value of which is less than a third digital threshold value (S3); - allocation (56) of a first group of calibrated pixels (G23) of the CMOS sensor (26) to the pixels of the filter (30) selecting said first wavelength (XI), said first group of calibrated pixels (G23) corresponding to the pixels common to the second group of pixels (G2) and to the third group of pixels (G3); - allocation (57) of a second group of calibrated pixels (G13) of the CMOS sensor (26) to the pixels of the filter (30) selecting said second wavelength (X2), said second group of calibrated pixels (G13) corresponding to the pixels common to the first group of pixels (G1) and to the third group of pixels (G3); - assignment (58) of a third group of calibrated pixels (G12) to the pixels of the filter (30) selecting said third wavelength (X3), said third group of calibrated pixels (G12) corresponding to the pixels common to the first group of pixels (G1) and to the second group of pixels (G2).

2. Image acquisition device (25) according to claim 1, in which the selective filter (30) is a filter integrating transparent pixels.

3. Image acquisition device (25) according to claim 2, in which the selective filter (30) is a SPARSE matrix type color filter comprising at least as many transparent pixels as wavelength selective pixels.

4. Image acquisition device (25) according to one of claims 2 or 3, in which the pixels of the CMOS sensor (26) not assigned to a wavelength following the calibration phase (Pc) are assigned to the radiation coming from the transparent pixels of the selective filter (30).

5. Image acquisition device (25) according to one of claims 1 to 4, wherein the image intensifier tube (13) comprises a microlens array (32) fixed to the wavelength-selective filter (30) so as to focus, at the output of the input window (15), the photons in the plane of the photocathode (16); the focusing of the photons being determined so as to anticipate the optical aberrations of the input window (15) and / or the dispersion of the electrons between the electron multiplier (18) and the phosphorescent screen (20) so as to avoid the overlap of the colors at the optical output interface (21).

6. Image acquisition device (25) according to claim 5, wherein the microlens array (32) has two, three or four microlenses per pixel of the wavelength-selective filter (30).

7. Image acquisition device (25) according to claim 5, wherein the microlens array (32) has between four and nine microlenses per pixel of the wavelength-selective filter (30).

8. Image acquisition device (25) according to one of claims 5 to 7, in which the pitch of the microlens array (32) is between 0.625 pm and 25 pm, preferably between 2 pm and 12 pm and preferentially between 4 pm and 8 pm.

9. Image acquisition device (25) according to one of claims 5 to 8, in which the ratio between the pitch defined by the distance between two consecutive pixels of the selective filter (30) and the pitch of the microlens array (32) is between 2 inclusive and 8 inclusive, preferably between 3 inclusive and 6 inclusive.

10. Image acquisition device (25) according to one of claims 1 to 9, in which the CMOS sensor (26) is only photosensitive.

11. Image acquisition device (25) according to one of claims 1 to 10, in which the phosphorescent screen (20) is based on P45 type phosphor.

12. Image acquisition device (25) according to one of claims 1 to 11, in which the selective filter (30) comprises a number of pixels between 20,000 and 8,000,000, preferably between 40,000 and 1,000,000.

13. Image acquisition device (25) according to one of claims 1 to 12, in which the pitch defined by the distance between two consecutive pixels of the selective filter (30) is between 5 pm and 50 pm, and preferably between 10 pm and 30 pm.

14. Image acquisition device (25) comprising: - an image intensifier tube (13) comprising the following elements fixed within an internal vacuum enclosure (24): ■ an input window (15) configured to receive and transmit photons; ■ a filter (30) selective for at least three distinct wavelengths (1, 2, 3), comprising a first set of pixels fixed on an external face of the input window (15); ■ a photocathode (16) capable of converting photons transmitted by the input window (15) into photoelectrons (28), said photocathode (16) being fixed on an internal face of the input window (15); ■ an electron multiplier (18) comprising two parallel faces (38, 39), capable of multiplying the photoelectrons (28) into secondary electrons (29); ■ a phosphorescent screen (20) transforming the secondary electrons (29) into photons; and ■ an optical output interface (21) configured to capture the photons emitted by the phosphorescent screen (20) to transmit them to the output of the image intensifier tube (13); - a CMOS sensor (26) having a sensitive area configured to transform photons transmitted by the optical output interface (21) into images and comprising a second set of pixels, each pixel of the second set of pixels of the CMOS sensor (26) being configured to be assigned to one or more pixels of the first set of pixels of the wavelength-selective filter (30), by means of a calibration phase (Pc); characterized in that said sensitive area of the CMOS sensor (26) has a number of pixels of between 8 million and 40 million; and in that said image acquisition device (25) comprises a control unit configured to implement a calibration phase (Pc) which comprises the following steps: - transmission (50) of radiation of a first wavelength (XI) on the input window (15), the first wavelength (XI) corresponding to one of said at least three wavelengths (XI, X2, X3) selected by said filter (30); - determination (51) of a first group of pixels (Gl), the least sensitive to said first wavelength (XI), the numerical value of which is less than a first threshold numerical value (SI); - transmission (52) of radiation of a second wavelength (X2) on the input window (15), the second wavelength (X2) corresponding to one of said at least three wavelengths (XI, X2, X3) selected by said filter (30); - determination (53) of a second group of pixels (G2), the least sensitive to said second wavelength (X2), the digital value of which is less than a second threshold digital value (S2); - transmission (54) of radiation of a third wavelength (X3) on the input window (15), the third wavelength (X3) corresponding to one of said at least three wavelengths (XI, X2, X3) selected by said filter (30); - determination (55) of a third group of pixels (G3), the least sensitive to said third wavelength (X3), the digital value of which is less than a third digital threshold value (S3); - allocation (56) of a first group of calibrated pixels (G23) of the CMOS sensor (26) to the pixels of the filter (30) selecting said first wavelength (XI), said first group of calibrated pixels (G23) corresponding to the pixels common to the second group of pixels (G2) and to the third group of pixels (G3); - allocation (57) of a second group of calibrated pixels (G13) of the CMOS sensor (26) to the pixels of the filter (30) selecting said second wavelength (2), said second group of calibrated pixels (G13) corresponding to the pixels common to the first group of pixels (G1) and to the third group of pixels (G3); - assignment (58) of a third group of calibrated pixels (G12) to the pixels of the filter (30) selecting said third wavelength (3), said third group of calibrated pixels (G12) corresponding to the pixels common to the first group of pixels (G1) and to the second group of pixels (G2).

15. Color night vision system (11) comprising: - a goal (12); - an image acquisition device (25) according to one of the preceding claims 1 to 14; and - a screen (14) for displaying the image of said CMOS sensor (26).

Images