EP3900320A1

EP3900320A1 - Image capture device and associated system for monitoring a driver

Info

Publication number: EP3900320A1
Application number: EP19791287.6A
Authority: EP
Inventors: Aleksandar Popovic
Original assignee: Valeo Comfort and Driving Assistance SAS
Current assignee: Valeo Comfort and Driving Assistance SAS
Priority date: 2018-12-19
Filing date: 2019-10-30
Publication date: 2021-10-27
Also published as: FR3091114B1; CN113302912A; US20220048386A1; US11845335B2; WO2020126179A1; FR3091114A1

Abstract

The invention relates to an image capture device (1) comprising an image sensor (9) making it possible to obtain both an infrared image (31) and an image in the visible range (35) by means of an array of optical filters comprising first filtering elements, having a transmission spectral band in the infrared range, as well as second filtering elements, having a transmission spectral band in the visible range. The image capture device further comprises: - a lighting device (11) configured to emit infrared radiation, and - a computer (13), programmed to control the power of the infrared radiation emitted by the lighting device, as a function of said ambient luminosity in the infrared range. The invention also relates to a monitoring system (2) for monitoring a driver (3), comprising such an image capture device.

Description

IMAGE CAPTURE DEVICE AND ASSOCIATED SYSTEM FOR

DRIVER MONITORING

[0001] The present invention relates to an image capture device. It also relates to a device for monitoring a driver of a vehicle comprising an image capture device.

Image capture devices for viewing a given scene, both in the visible and in the infrared domain, and this with the same image sensor, have been developed recently.

The image sensor of such a device, somehow hybrid, is sometimes called an "RGB-IR" image sensor (according to the acronym of Red Green Blue - InfraRed, that is- i.e. Red Green Blue - InfraRed).

This image sensor comprises a matrix of photosensitive pixels and a network of elementary optical filters coinciding with these different photosensitive pixels.

Part of these elementary optical filters are color filters. They each transmit part of the visible radiation received by the image sensor, this part corresponding in practice to a red, a green, or a blue. These various elementary color filters make it possible to acquire a color image, for example of the “RGB” type.

The other elementary optical filters of the image sensor are transparent, at least in part, in the infrared domain. They make it possible to acquire an image of the scene in question in the infrared domain.

The color image and the image in the infrared domain, which contain additional information, are thus obtained with the same sensor, which is advantageous in particular in terms of cost and size.

However, the ambient brightness in the environment of an image capture device is often clearly different in the visible range and in the infrared range. With such an image sensor, it is therefore generally not possible to obtain an optimal exposure both for the color image and for the image in the infrared range.

In this context, the present invention provides an image capture device comprising an image sensor, which comprises: a network of optical filters receiving electromagnetic radiation and comprising first filter elements each capable of transmitting a first part of the electromagnetic radiation included in a given wavelength range of infrared, as well as second filter elements each capable of transmit at least one component of a second part of the electromagnetic radiation located in the visible, and

a matrix of photosensitive pixels comprising first photosensitive pixels arranged so as to capture the first part of the electromagnetic radiation transmitted by the first filtering elements, as well as second photosensitive pixels arranged so as to capture the component transmitted by the second filtering elements, each first and second photosensitive pixels being able to generate an electrical signal representative of the power of the electromagnetic radiation which it picks up.

According to the invention, the image capture device further comprises:

a lighting device configured to emit infrared radiation in a field of vision of the image capture device, said infrared radiation being located at least partly in said range of wavelengths transmitted by the first filter elements, and

- a computer, programmed to execute the following steps: a) determining an ambient brightness in the infrared range, b) controlling the power of the infrared radiation emitted by the lighting device, as a function of said ambient brightness in the range infrared, c) acquire the electrical signals generated by the first and second photosensitive pixels, compose a first image from the electrical signals generated by the first photosensitive pixels, and compose a second image from the electrical signals generated by the second photosensitive pixels.

Infrared radiation, which illuminates the scene located in the field of vision of the image capture device, includes the radiation emitted by the lighting device, as well as the infrared radiation possibly coming from other surrounding sources (light solar for example). The lighting device, controlled as a function of the ambient light in the infrared domain, makes it possible to control the total power of the infrared radiation which illuminates the scene.

This allows optimal exposure of the first image (“infrared” image), even if the exposure time (that is to say the integration time), or possibly other exposure parameters (gain, aperture), are imposed elsewhere, for example to obtain an optimal exposure of the second image (image in the visible range).

The computer can also be programmed to execute the following steps: a ') determining an ambient brightness in the visible range, and b') controlling at least one of the following exposure parameters as a function of said ambient brightness in the visible range:

an integration duration, each of said electrical signals being representative of an electric charge or voltage, accumulated by the corresponding photosensitive pixel during said integration duration,

- a gain applied to said electrical signals,

- an aperture of a diaphragm of an optical system of the image capture device.

Thus controlling one or more of these exposure parameters makes it possible to obtain a suitable exposure of the second image, and thus to prevent it from being overexposed or underexposed. As explained above, controlling the power of the infrared radiation emitted by the lighting device also makes it possible to obtain a suitable exposure of the first image (“infrared” image), even if the exposure parameters mentioned above are also fixed, depending on the ambient brightness in the visible range.

In a possible embodiment, it can be provided that the computer is programmed for, step c) having been executed previously, determining said ambient brightness in the visible range, in step a '), depending on brightness values of at least part of the image pixels of the second image which was produced during said previous execution of step c).

In a possible embodiment, the computer is programmed to, in step a '), determine said ambient brightness in the visible range so that that it is representative of a second level of average brightness, in the second image produced during said previous execution of step c).

In a possible embodiment, the computer is programmed to, in step b '), correct at least one of said exposure parameters as a function of a difference between, on the one hand, a second target value of average brightness, and, on the other hand, the second level of average brightness in the second image produced during the previous execution of step c).

In a possible embodiment, the computer is programmed for, step c) having been executed previously, determining said ambient brightness in the infrared range, in step a), as a function of values of brightness of at least part of the image pixels of the first image which was produced during said previous execution of step c).

In a possible embodiment, the computer is programmed to, in step a), determine said ambient brightness in the infrared range so that it is representative of a first brightness level average, in the first image produced during said previous execution of step c).

In a possible embodiment, the computer is programmed to, in step b), control the power of the infrared radiation emitted by the lighting device as a function of a difference between, on the one hand, a first target value of average brightness, and, on the other hand, the first level of average brightness in the first image produced during the previous execution of step c).

In a possible embodiment, said range of wavelengths transmitted by the first filter elements is between 700 nanometers and 1100 nanometers.

In a possible embodiment, the second filter elements (include red filter elements having a red bandwidth transmitting wavelengths at least between 550 nanometers and 700 nanometers, green filter elements have a green bandwidth transmitting wavelengths between at least 450 nanometers and 650 nanometers and blue filter elements have blue bandwidth transmitting wavelengths for example between 400 nanometers and 550 nanometers. Of course, the different characteristics, variants and embodiments of the invention can be associated with each other in various combinations as long as they are not incompatible or mutually exclusive of each other.

The invention also provides a system for monitoring a driver of a vehicle comprising an image capture device as described above, as well as a processing unit programmed to determine a level of incapacity driving the driver from at least said first image.

The level of unfit driving can include a level of drowsiness and / or a level of distraction of the driver.

In addition, various other characteristics of the invention appear from the attached description made with reference to the drawings which illustrate non-limiting forms of embodiment of the invention and where:

[Fig. 1] schematically represents a motor vehicle comprising an image capture device implementing the teachings of the invention, seen from the side,

[Fig. 2] shows in more detail certain elements of the image capture device of FIG. 1,

[Fig. 3] schematically represents a network of optical filters with which an image sensor of the image capture device of FIG. 1 is provided, seen from the front,

[Fig. 4] schematically represents this same image sensor, seen from the side,

[Fig. 5] schematically represents an infrared image and a color image delivered by the image capture device of FIG. 1, and

[Fig. 6] schematically represents steps executed by a computer of the

image capture device of figure 1.

Figure 1 shows a vehicle 5, here a motor vehicle, provided with a monitoring system 2 of a driver 3 of the vehicle. This monitoring system 2 comprises an image capture device 1 and a processing unit 20 which is programmed to determine a level of unsuitability for driving the driver 3, l _L , from one or more images delivered by the image capture device 1.

The image capture device 1 comprises an image sensor 9 (Figure 2) and an optical system 10 such as a lens. The optical system 10 forms, on the sensor of images 9, an image of the content of the field of view 14 of the image capture device 1.

As shown in Figure 1, the image capture device 1 is located in a passenger compartment 7 of the vehicle, in a region adjacent to a vehicle windshield. It is for example integrated in a dashboard, in a dashboard or in a vehicle control console.

The image capture device 1 is oriented so that its field of vision 14 covers the area usually occupied by the head of the driver 3 when the latter is seated in the driver's seat. The image capture device 1 can thus capture images of the driver's face 3.

Remarkably, the image sensor 9 is a somewhat hybrid sensor, making it possible to visualize the content of the field of vision 14 both in the infrared domain, and in the visible domain. In this case, the image sensor 9 is provided with a particular network 170 of optical filters (FIG. 3) which allows, with the same matrix 19 of photosensitive pixels, to acquire both:

a first image 31, hereinafter called “infrared image”, produced from a first part of the electromagnetic radiation collected by the optical system 10, this first part being located in the infrared domain, and

- A second image 35, produced from a second part of the electromagnetic radiation collected by the optical system 10, this second part being located in the visible range.

In the example described here, this second image 35 is a color image (it is also called "color image" in the following).

Thanks to this particular image sensor 9, the image capture device 1 is able to capture the infrared image in question, as well as this color image, while remaining very compact. The infrared image 31 can for example be used to monitor the driver 3 and determine his level of inability to drive l _L. The color image is more pleasing to the human eye and contains more information than an infrared image (usually displayed in grayscale). The color image can therefore be acquired for multiple purposes, for example it can be used to communicate with a remote electronic device, for example in the context of a teleconference, or even be kept in memory for security reasons or as a "souvenir photo" immortalizing a journey.

The image capture device 1 also includes a computer 13, comprising at least one processor and an electronic memory, programmed to control one or more exposure parameters, such as the integration time and the gain used for obtain the images 31, 35 in question. The computer 13 is programmed to control this or these exposure parameters as a function of an ambient brightness in the visible range. This makes it possible to obtain a suitable exposure of the second image 35.

The image capture device 1 also includes a lighting device 11, for

illuminate the field of vision 14 in the infrared domain. This lighting device 11 is controlled by the computer 13 as a function of an ambient brightness in the infrared range. Thus controlling the power of the infrared radiation emitted by the lighting device 11 also makes it possible to obtain a suitable exposure of the infrared image, even if the exposure parameter or parameters mentioned above are adjusted as a function of the ambient brightness. in the visible domain, not in the infrared domain.

The lighting device 11 controlled by the computer 13 therefore makes it possible, in a way, to avoid too great a difference between the ambient brightness in the visible range and that in the infrared range, which allows , for the same integration time ti or for the same gain G, to obtain an appropriate exposure both for the infrared image 31 and for the color image 35.

The various elements and functionalities of this image capture device 1 can now be described in more detail. The image sensor 9 will be described in a first. The way to control the exposure parameters (integration time ti, gain G, aperture Ap) will be presented in a second step. The lighting device 11 and its control will be presented next, before describing the way of determining the driver unsuitability parameter 3.

RGV-IR image sensor.

As can be seen in Figures 3 and 4, the network of optical filters 170 of the image sensor 9 is arranged opposite the matrix 19 of photosensitive pixels of this sensor, so as to filter electromagnetic radiation from the field of vision 14 of the image capture device 1 (radiation which has been collected by the optical system 10), before this radiation reaches the photosensitive pixels 21, 2 in question.

The network of optical filters 170 comprises several filtering elements 171, 172, 173, 174, that is to say several elementary optical filters, each arranged opposite one of the photosensitive pixels 21, 21 '. Each photosensitive pixel thus captures part of the electromagnetic radiation which has been filtered by the filter element with which it is associated. As explained below, these individual filter elements 171, 172, 173, 174 are of different types, for example of the blue, green, red and infrared type, which makes it possible to acquire the color image and the infrared image mentioned above.

In this case, the network of optical filters 170 comprises:

first filter elements 171, each capable of transmitting a first part of the electromagnetic radiation situated in a given wavelength range of the infrared, and

- Second filter elements 172, 173, 174, each capable of transmitting at least one component of a second part of the electromagnetic radiation located in the visible range, between 400 nanometers and 700 nanometers.

In the embodiment described here, the first filter elements 171 transmit only wavelengths located in said infrared wavelength range. This wavelength range mainly extends beyond 700 nanometers. It can for example extend from 700 nanometers to 1100 nanometers.

In the example described here, the second filter elements 172, 173, 174, transmit only the wavelengths between 400 and 700 nanometers. As a variant, they could however transmit both wavelengths located in the visible range, and wavelengths located in the infrared range.

The second filter elements 172, 173, 174 here include red filter elements 172, green filter elements 173 and blue filter elements 174. The terms "red", "green", "blue" are used in their sense common. The values of the red, green and blue bandwidths set out below are given by way of non-limiting example.

The red filter elements 172 have a red bandwidth transmitting the component of the second part of the electromagnetic radiation having wavelengths comprised for example mainly between 550 nm and 700 nm. The green filter elements 173 have a green bandwidth transmitting the component of the second part of the electromagnetic radiation having wavelengths comprised for example mainly between 450 nm and 650 nm.

The blue filter elements 174 have a blue passband transmitting the component of the second part of the electromagnetic radiation having wavelengths comprised for example mainly between 400 nm and 550 nm.

The various filter elements 171, 172, 173, 174 of the network of optical filters are arranged with respect to each other so as to form a pattern 175, repeated regularly to form the network of optical filters 170. This pattern 175 comprises here four filter elements, in this case: one of the first filter elements 171, one of the red filter elements 172, one of the green filter elements 173, and one of the blue filter elements 174. These four elements adjacent filters form a square. As shown in the figures, the network of optical filters 170 is therefore comparable to a so-called “Bayer” network, in which one of the green filter elements has been replaced by a transparent element in the infrared. As a variant, the pattern of filter elements, repeated several times to form the network of optical filters, could be constituted differently (by including, for example, more green filter elements than red or blue filter elements).

As regards the matrix 19 of photosensitive pixels, it comprises:

first photosensitive pixels 21, arranged so as to capture the first part of the electromagnetic radiation transmitted by the first filter elements 171, and

- second photosensitive pixels 2, arranged so as to capture the different components, red, green and blue, transmitted by the second filter elements 172, 173, 174.

Each of the first and second photosensitive pixels 21, 2 produces, by electrical photo effect, an electrical signal representative of the power of the electromagnetic radiation that it has captured (in a way, each photosensitive element behaves like a photodiode ). This electrical signal is produced in the form of a charge or an electrical voltage presented between two terminals of the electrical capacity that constitutes the photosensitive element. This electrical signal is produced by the photosensitive pixel 21, 21 'considered at the end of a given integration time ti. More specifically, before acquiring the first and second images 31, 35, the electric charges, or the electric voltages between the terminals of the first and second photosensitive pixels 21, 2 are reset (reset to zero), then, each of these photosensitive pixels 21, 2 accumulates electrical charges, produced by photoelectric effect, during said integration time ti. The electrical signal produced by the photosensitive pixel 21, 2 considered corresponds to the charge, or to the electrical voltage between the terminals of this photosensitive pixel 21, 2, at the end of this integration time ti, sometimes called exposure time, acquisition time, or integration time. The electrical signal produced by each of the first and second photosensitive pixels 21, 21 'is representative of the number of photons received by the photosensitive pixel considered during the integration time ti (for example proportional to this number of photons).

In the embodiment described here, the semiconductor substrate of the matrix 19 of photosensitive pixels 21, 2 is made of silicon (appropriately doped). The sensitivity of the photosensitive pixels in the infrared domain is thus limited to the near infrared domain: the sensitivity range of the first photosensitive pixels 21, provided with the first filter elements 171, is here between 700 nanometers and 1100 nanometers.

The image sensor 9 is for example of the CMOS type (according to the English acronym of

“Complementary Metal Oxide Semiconductor”) or CCD type (according to the English acronym for “Charge Coupled Device”).

G gain is applied to the various electronic signals generated by the first and second photosensitive pixels 21, 21 '. Each of these signals is multiplied by this gain G to obtain a corresponding amplified electrical signal, which is for example then digitized by means of an analog-digital converter. In the case of a CCD type sensor, one can for example provide that the same amplifier applies the aforementioned gain G to the various electrical signals, previously generated by the photosensitive pixels 21, 2 then transferred line by line and column by column ( for example) to this amplifier. In the case of a CMOS sensor, each photosensitive pixel 21, 2 can be provided with an amplifier of its own, these different amplifiers each applying the same gain G, to the corresponding electrical signal. As a variant, these different amplifiers could apply different gains, friends whose average value is equal to the aforementioned gain G. A picture taken by the image sensor 9 comprises:

the exposure of the first and second photosensitive pixels 21, 2, during the integration time ti, to generate said electrical signals, and

- the subsequent amplification of these electrical signals, by gain G.

During this shooting, the values of the integration time ti and the gain G are fixed by the computer 13, which controls the image sensor 9.

The amplified electrical signals produced during this shooting are processed by the computer 13, during a step c) (FIG. 6), to produce the infrared image 31 and the color image 35.

Composition of the infrared image and the color image

The infrared image 31 is formed of a matrix of image pixels 33 (FIG. 5), associated with the different photosensitive pixels 21, 21 ′ of the image sensor 9. In the embodiment described here, the computer 13 is programmed to implement an interpolation algorithm making it possible to constitute a “complete” infrared image 31 although only one photosensitive pixel 21 out of four captures infrared radiation in the image sensor. In other words, the infrared image 31 here comprises as many image pixels 33 as what the image sensor 9 comprises of photosensitive pixels, first 21 and second 21 'photosensitive pixels included.

Each image pixel 33 of the infrared image 31 is associated with a brightness value, representative of the value of one or more of the amplified electrical signals originating from the first photosensitive pixels 21. Here, this brightness value is determined by function of the values of the amplified electrical signals coming from the first photosensitive pixels 21 which are closest to the position corresponds, on the image sensor 9, to the image pixel 33 considered.

The computer 13 is also programmed to compose the color image 35, shown diagrammatically in FIG. 5, from the amplified electrical signals coming from the second photosensitive pixels 2.

The color image 35 is also formed from a matrix of image pixels (not shown), associated with the different photosensitive pixels 21, 21 ′ of the image sensor 9. As for the infrared image, the computer 13 is programmed here to implement an interpolation algorithm making it possible to constitute a “complete” color image 35 although only one photosensitive pixel 21 out of four captures radiation located in the red, or green, or blue bandwidth mentioned above. Otherwise formulated, the color image here comprises as many image pixels as what the image sensor 9 comprises of photosensitive pixels, first 21 and second 21 'photosensitive pixels included.

Each pixel image of the color image 35 is associated with a brightness value, representative of values of some of the amplified electrical signals from second photosensitive pixels 21 '. This brightness value is representative of the intensity of the visible electromagnetic radiation received by the second photosensitive pixels 2 which, on the image sensor 9, are located in the immediate vicinity of the position which, on this sensor, is associated with the image pixel considered.

In this case, the computer 13 is programmed here to initially compose three monochrome channels based on the amplified electrical signals from the second photosensitive pixels 2. The computer 13 thus composes a red channel 37, from the amplified electric signals coming from the second photosensitive pixels located opposite the red filter elements 172. It also composes a green channel 39 from the amplified electric signals coming from the second photosensitive pixels located opposite the green filter elements 173. Finally, it composes a blue channel 41 from the amplified electrical signals coming from the second photosensitive pixels associated with the blue filter elements 174.

Each of these channels, red 37, green 39 and blue 41, is a gray level image (each pixel image of the channel considered has a brightness value, but no hue or chrominance value), associated with the color of the channel considered and of the same size as the color image 35 (that is to say comprising the same number of image pixels).

To develop the color image 35, the computer 13 then implements a

algorithm making it possible to combine the brightness values of the red channel 37, the green channel 39 and the blue channel 41, for each image pixel, in order to obtain, for the image pixel "in color" considered:

a global luminosity value for the visible range (for example equal to the average of the luminosities of the different channels), and

- two chrominance values, complementary, representative of the chromatic properties of the captured light, at the point considered of the image sensor 9.

As shown in Figure 6, the computer 13 is programmed here to execute step c) several times in succession, during which it acquires the signals amplified electrics delivered by the image sensor 9 then composes the infrared image 31 and the color image 35. This step is carried out here by a pretreatment module 130 of the computer 13 (FIG. 2).

Control of exposure parameters

The exposure parameters of the image capture device 1, namely the integration time ti, the gain G, and, optionally, the aperture Ap of a diaphragm 12 of the optical system 10, are controlled here by an exposure control module 132 of the computer 13 (Figure 2), during steps a ') and b') (Figure 6) described below.

The exposure control module 132 is programmed to determine the ambient brightness in the visible range which was mentioned above, during step a ').

The ambient brightness in the visible range is representative of the power of visible electromagnetic radiation (the components of which are mainly between 400 nanometers and 700 nanometers), coming from an environment of the image capture device 1 , here coming from the field of vision 14 of this device, and received by a light sensor. In the embodiment described here, this brightness sensor is produced by means of the image sensor 9 which has been described above.

Alternatively, it could however be a separate sensor from the image sensor, such as a photodiode arranged so as to receive radiation from the field of vision of the image capture device.

Here, the ambient brightness in the visible range is determined from a color image 35, delivered by the pretreatment module 130 after a previous execution of step c), and corresponding therefore to a shooting previously carried out by the image sensor 9.

In this case, the exposure control module 132 is programmed to determine the ambient brightness in the visible range as a function of the brightness values of at least part of the image pixels of this color image 35, previously acquired by the image capture device 1.

The ambient brightness in the visible range can, as here, be determined by calculating an average of the brightness values of the image pixels of this color image 35. The ambient brightness in the visible range is then representative an average brightness level in this color image 35, hereinafter called the second brightness level, and noted L _v .

The average in question can relate to all of the image pixels of the color image 35, or relate only to some of these image pixels, located in an area of interest of the image corresponding for example to l image of the driver's face 3. Furthermore, it can be provided that this average takes into account only the brightness values of the image pixels which satisfy a given criterion, for example which are included in a given range of values. Thus, the second level of luminosity L _v can, by way of example, be representative of the average luminosity of the zones of low luminosity of the color image 35, or of the zones of intermediate luminosity of this image.

As a variant, the ambient brightness in the visible range could be determined as a function of the brightness values of the various chromatic channels of the color image considered (red, green and blue channel), possibly affected by different weighting coefficients , instead of being determined as a function of luminosity values of the image pixels of the overall color image resulting from the fusion of these three chromatic channels.

In step b '), the exposure control module 132 controls the aforementioned exposure parameters, as a function of the ambient brightness in the visible range determined in step a') above.

For this, the exposure control module 132 here corrects the values of the exposure parameters, which are then applied to the image sensor 9 and to the diaphragm 12, as a function of a difference s2 between:

on the one hand, a target value of average luminosity, hereinafter called the second target value L _v, o, and

- on the other hand, the second level of brightness L _v (average brightness in the color image 35 acquired previously).

This correction is made so as to gradually bring the second brightness level L _v to the second target value L _v, o, during repetitions of steps a '), b') and c). This correction can consist, for example, of adding a corrective term to a previous value of the exposure parameter considered, this corrective term being proportional to the difference s2 mentioned above (proportional correction). More generally, this correction consists in slaving the second level of brightness L _v to the second target value Lv _, o. This control can in particular be of the proportional, proportional-integral, or proportional, integral and derived type (that is to say of the "PID" type).

The second target value L _v, o corresponds for example to the average brightness in an image considered to be adequately exposed.

An image is considered to be suitably exposed when, for example:

the average brightness in this image is included in a given interval, this interval extending for example from a quarter to three quarters of the maximum brightness value which can be associated with an image pixel,

- And / or when the proportion of image pixels in the image, whose brightness value is equal to said maximum brightness value (image pixels in high saturation), is less than a given threshold, this threshold being for example equal to A quarter,

- And / or when the proportion of image pixels in the image, whose brightness value is equal to a minimum brightness value (image pixels in low saturation), is less than a given threshold, this threshold being for example equal to A quarter.

The second target value L _v, o can for example be between a quarter and three quarter of the maximum brightness value mentioned above, or between a third and two thirds of this maximum value. For example, if the brightness values in question are coded on eight bits, being between 0 and 255, the maximum brightness value in question (high saturation) is equal to 255, and the second target value L _{v ,} o can then be between 63 and 191, for example, or between 85 and 170.

In the example described here, the exposure control module 132 controls both the integration time ti, the gain G, and the aperture Ap of the diaphragm of the optical system 10 (aperture diaphragm) . Alternatively, the opening Ap could however be fixed (or possibly be manually adjustable). Similarly, the exposure control module 132 could be programmed for, the integration time ti being fixed, controlling only the value of the gain G, as a function of the second level of brightness L _v in the color image 35 ( or, conversely, to control only the integration time ti). Such G gain control is sometimes called "automatic gain control" or "AGC" in the specialized literature.

Lighting device The lighting device 11 is capable of emitting infrared radiation located at least partly in the range of wavelengths transmitted by the first filter elements 171 of the optical filter network of the image sensor 9 (range including it is recalled that it extends here from 700 nanometers to 1100 nanometers). The lighting device 11 can for example be produced by means of light-emitting diodes.

The infrared radiation emitted by the lighting device 11 is emitted in the form of a light beam 15, directed so as to illuminate at least part of the field of vision 14 of the image capture device 1 (FIG. 1). In this case, this light beam is directed towards the area usually occupied by the face of the driver 3, when seated in the driver's seat.

The power of the infrared radiation emitted by the lighting device 11 is controlled by a lighting control module 131 of the computer 13 (Figure 2). For this, this module here controls the electrical power P _E which supplies the lighting device 11.

The lighting control module 131 is programmed more precisely to control the power of the infrared radiation emitted as a function of ambient brightness in the infrared domain.

The ambient brightness in the infrared range is defined here in a manner comparable to the ambient brightness in the visible range presented above, but for the infrared range. The control of the power of the infrared radiation emitted is comparable to the control of the exposure parameters mentioned above, but it is carried out according to the ambient light in the infrared domain, instead of being carried out according to the ambient brightness in the visible range.

The lighting control module 131 is programmed to control the power of the infrared radiation emitted during steps a) and b), shown in Figure 6 and described below.

During step a), the lighting control module 131 determines the ambient brightness in the infrared range.

The ambient brightness in the infrared domain is representative of the power of infrared electromagnetic radiation (whose components extend beyond 700 nanometers), coming from the environment of the capture device. images 1, here coming from the field of vision 14 of this device, and received by a brightness sensor. In the embodiment described here, this brightness sensor is produced by means of the image sensor 9.

Alternatively, it could however be a separate sensor from the image sensor, such as an infrared photodiode, arranged to receive radiation from the field of vision of the image capture device.

Here, the ambient brightness in the visible range is determined from an infrared image 31, delivered by the pretreatment module 130 after a previous execution of step c), and therefore corresponding to a shot previously made by the image sensor 9.

In this case, the lighting control module 131 is programmed to determine the ambient brightness in the infrared domain as a function of the brightness values of at least part of the image pixels 33 of this infrared image. 31, previously acquired by the image capture device 1.

The ambient brightness in the infrared range can, as here, be determined by calculating an average of the brightness values of the image pixels 33 of this infrared image 31. The ambient brightness in the infrared range is then representative of an average brightness level in this infrared image 31, hereinafter called the first brightness level, and denoted L _{| R.}

The average in question can relate to all of the image pixels 33 of the infrared image 31, or relate only to some of these image pixels 33, located in an area of interest of the image corresponding for example to the image of the driver's face 3. Furthermore, it is possible to provide for this average to take into account only the brightness values of the image pixels 33 which satisfy a given criterion, for example which are included in a given range of values. Thus, the first brightness level L _{| R} can, by way of example, be representative of the average brightness of the low light areas of the infrared image 31, or of the intermediate light areas of this image.

In step b), the lighting control module 131 controls the electrical power P _E which supplies the lighting device 11, as a function of the ambient brightness in the infrared range determined in step a) former.

For this, the lighting control module 131 here corrects the value of the electric power P _E , as a function of a difference if between: - on the one hand, a target value of average brightness in the infrared range, hereinafter called the first target value L _{| R,} o, and

- on the other hand, the first brightness level L _{| R} (average brightness in the infrared image 31 acquired previously).

This correction is made so as to gradually bring the first brightness level L _{| R} to the first target value L _IR.O , during repetitions of steps a), b) and c). This correction can consist, for example, of adding a corrective term to a previous value of the electrical power P _E , this corrective term being proportional to the difference if mentioned above (proportional correction). More generally, this correction consists in slaving the first level of brightness U _R to the first target value L _{| R,} o. This control can in particular be of the proportional, proportional-integral, or proportional, integral and derived type (that is to say of the "PID" type).

The first target value L _{! R,} o corresponds for example to the average brightness in an image considered to be adequately exposed (an example of definition of an image considered to be adequately exposed was given above).

The first target value L _{! R,} o can for example be between a quarter and three quarter of the maximum brightness value that can be associated with an image pixel 33 of the infrared image 31, or be between a third and two thirds of this maximum value. For example, if the brightness values of the image pixels 33 are coded on eight bits, being between 0 and 255, the maximum brightness value mentioned above (high saturation) is equal to 255, and the first target value L _{! R,} o can then be between 63 and 191, for example, or between 85 and 170.

Driver monitoring system

The electronic processing unit 20 of the monitoring system 2 is programmed to determine the level of inability to drive l _L of the driver 3 from at least one of the infrared images 31 produced by the capture device. of images 1. The level of inability to drive l _L comprises for example a level of drowsiness of the driver 3 and / or a level of distraction of the driver 3 (the level of inability to drive l _L can in particular be driver drowsiness level 3 or distraction level).

The processing unit 20 can for example be programmed so as to analyze the infrared image 31 in question, or a sequence of infrared images 31 produced by the image capture device 1, in order to identify the face of the driver 3 and / or certain areas of the face of the driver 3, in particular the areas of the infrared image 31 corresponding to the eyes of the driver 3. The processing unit 20 can then determine the level of drowsiness of the driver 3 by measuring the duration and / or the blinking frequency of the eyes of the driver 3, previously identified in the infrared image 31.

The processing unit 20 can determine the distraction level of the driver 3 as a function of a posture of the head of the driver 3 deduced from the infrared image 31, and as a function of the evolution of this posture during time.

The processing unit 20 can also evaluate (by analysis of the infrared image 31, or of a sequence of infrared images 31), and use to determine the level of distraction and / or the level of drowsiness, the viewing direction of the driver 3 or the evolution of this viewing direction over time.

The processing unit 20 can also evaluate (by analysis of the infrared image 31, or of a sequence of infrared images 31), and use to determine the level of distraction and / or the level of drowsiness, the diameter of the pupil of at least one eye of the conductor 3 (and precisely the variations of this diameter).

The processing unit 20 can be programmed to, when determining the level of inability to drive l _L of the driver, also take account of one or more color images 35 delivered by the capture device d 'images 1.

The color image 35 can be used in other applications.

The computer 13 can for example transmit the color image 35, or a sequence of color images 35 to a telecommunication module 43 of the vehicle 5. This telecommunication module 43 is configured to transmit the color image 35, or the sequence of color images 35 received at a remote electronic device, for example a multifunction mobile or a computer, for example via a wifi transmitter. The color image 35 or the color image sequence 35 can then be used within the framework of a teleconference, for example a videoconference.

The computer 13 could also transmit the color image 35, or the sequence of color images 35 to a memory of the vehicle 5 so that it is stored there.

Different variants can be made to the image capture device or to the driver monitoring system which have been described above. For example, the second image, composed from the electrical signals generated by the second photosensitive pixels of the image sensor, could be a monochrome image, instead of being a color image. The second filter elements of the network of optical filters could moreover all be of the same type (for example being all of the green filter elements), instead of comprising three different types of optical filters (respectively red, green and blue).

On the other hand, the different functionalities of the computer could be distributed differently between modules. A greater number of modules could be used, or, on the contrary, the same module could execute all of the operations, executed by the computer, which have been described above. It is noted that the term module can designate an electronic circuit, or a part of electronic circuit distinct from the other modules, or a specific group of instructions stored in the memory of the computer.

Furthermore, the first and second images mentioned above could correspond to somewhat raw images (sometimes called “RAW” images in English), obtained without interpolation. In this case, the number of image pixels of the infrared image, for example, would be equal to the number of said first photosensitive pixels, instead of being equal to the total number of photosensitive pixels of the image sensor.

Claims

[Claim 1] [Image capture device (1) comprising an image sensor (9), which

includes:

- an array of optical filters (170) receiving radiation

electromagnetic and comprising first filter elements (171) each capable of transmitting a first part of the electromagnetic radiation included in a given wavelength range of infrared, as well as second filter elements (172, 173, 174) each capable transmitting at least one component of a second part of the electromagnetic radiation located in the visible, and

- a matrix (19) of photosensitive pixels comprising first photosensitive pixels (21) arranged so as to capture the first part of the electromagnetic radiation transmitted by the first filter elements (171), as well as second photosensitive pixels (2) arranged picking up the component transmitted by the second filter elements (172, 173, 174), each of the first and second photosensitive pixels (21, 21 ') being capable of generating an electrical signal representative of the power of the electromagnetic radiation that it picks up,

the image capture device (1) being characterized in that it further comprises:

- a lighting device (11) configured to emit infrared radiation in a field of vision (14) of the image capture device (1), said infrared radiation being located at least partially in said range of lengths of wave transmitted by the first filter elements (171), and

- a computer (13), programmed to execute the following steps:

a) determine an ambient brightness in the infrared range,

b) controlling the power of the infrared radiation emitted by the lighting device (11), as a function of said ambient brightness in the infrared domain,

c) acquiring the electrical signals generated by the first and second photosensitive pixels (21, 2), composing a first image (31) from the electrical signals generated by the first photosensitive pixels (21), and composing a second image (35) from the electrical signals generated by the second photosensitive pixels (2).

[Claim 2] An image capture device (1) according to claim 1, wherein the

computer (13) is programmed to execute the following steps:

a ') determine an ambient luminosity in the visible range, and

b ') control at least one of the following exposure parameters as a function of said ambient brightness in the visible range:

an integration duration (ti), each of said electrical signals being representative of an electric charge or voltage, accumulated by the photosensitive pixel (21, 2) corresponding during said integration duration (ti),

- a gain (G) applied to said electrical signals,

- An opening (Ap) of a diaphragm (12) of an optical system (10) of the image capture device (1).

[Claim 3] An image capture device (1) according to claim 2, wherein the

computer (13) is programmed for, step c) having been executed previously, determining said ambient brightness in the visible range, in step a '), as a function of brightness values of at least part of the pixels images of the second image (35) which was produced during said previous execution of step c).

[Claim 4] An image capture device (1) according to claim 3, wherein the

computer (13) is programmed to, in step a '), determine said ambient brightness in the visible range so that it is representative of a second average brightness level (L _v ), in the second image (35) produced during said previous execution of step c).

[Claim 5] An image capture device (1) according to claim 4, wherein the

computer (13) is programmed to, in step b '), correct at least one of said exposure parameters (ti, G, Ap) as a function of a difference between, on the one hand, a second value target (L _v, o) of average brightness, and, on the other hand, the second level of brightness (L _v ) average in the second image (35) produced during the previous execution of step c).

[Claim 6] Image capture device (1) according to one of claims 1 to 5, in which the computer (13) is programmed for, step c) having been executed

previously, determine said ambient brightness in the infrared range, in step a), as a function of brightness values of a part at minus the image pixels (33) of the first image (31) which was produced during said previous execution of step c).

[Claim 7] An image capture device (1) according to claim 6, wherein the

computer (13) is programmed to, in step a), determine said ambient brightness in the infrared range so that it is

representative of a first average brightness level (L _{| R} ), in the first image (31) produced during said previous execution of step c).

[Claim 8] An image capture device (1) according to claim 7, wherein the

computer (13) is programmed to, in step b), control the power of the infrared radiation emitted by the lighting device (11) as a function of a difference between, on the one hand, a first target value (L _{! R,} o) of average brightness, and, on the other hand, the first level of average brightness (L _{| R} ) in the first image (31) produced during the previous execution of step c).

[Claim 9] Image capture device (1) according to one of claims 1 to 8, in which said range of wavelengths transmitted by the first filter elements (171) is between 700 nanometers and 1100 nanometers.

[Claim 10] Image capture device (1) according to one of claims 1 to 9, in which the second filter elements (172, 173, 174) comprise red filter elements (172) having a red passband transmitting wavelengths between at least 550 nanometers and 700 nanometers, green filter elements (173) have a green bandwidth transmitting wavelengths at least between 450 nanometers and 650 nanometers and blue filter elements (174) have blue bandwidth transmitting wavelengths for example between 400 nanometers and 550 nanometers.

[Claim 11] Monitoring system (2) of a driver (3) of a vehicle (5) comprising an image capture device (1) according to one of claims 1 to 10, as well as a unit processing element (20) programmed to determine a level of inability to drive (l _L ) of the driver from said first image (31) at least.

[Claim 12] The monitoring system of claim 11, wherein the level

inability to drive (l _L ) includes a level of drowsiness or a level of driver distraction.