Vehicle occupant detection device and vehicle

Abstract

The present invention relates to a vehicle occupant detection device, comprising at least one infrared light source (1), a camera (2), and a computing unit (3) for controlling the infrared light source (1) and evaluating the camera image generated by the camera (2), wherein the infrared light source (1) and the camera (2) are directed toward an observation area (4), the infrared light source (1) is configured to illuminate at least one facial area of ​​a person (5) in the observation area (4), the camera (2) is configured to detect at least the illuminated facial area, and the computing unit (3) is configured to recognize the gaze direction of the person (5) in the camera image. The vehicle occupant detection device according to the present invention is characterized in that the infrared light source (1) is incorporated into a display device (6) and formed by a plurality of infrared light-emitting elements (7) arranged in a matrix, the computing unit (3) is further configured to activate IR light-emitting elements (7) selected to emit infrared light, and the computing unit (3) determines which IR light-emitting elements (7) to be activated according to the evaluation result of the camera image.

Description

Technical Field

[0001] The present invention relates to a vehicle occupant detection device of the kind defined in detail in the preamble of claim 1 and to a vehicle equipped with such a vehicle occupant detection device.

Background Art

[0002] Recent vehicles are equipped with various sensor systems for detecting the vehicle's surrounding environment and the interior of the vehicle. This enables the provision of various driver assistance systems for enhancing user comfort and improving road traffic safety. For example, it is well known to detect a vehicle driver using a camera. This makes it possible to recognize gestures and interpret them for input of operation actions, confirm the presence of vehicle occupants, identify vehicle occupants based on biometric features, or confirm the attention, health condition and / or distraction level of the vehicle driver.

[0003] Observing the eyes of a vehicle driver with a camera is particularly important. For example, the frequency of blinking can be a sign indicating the driver's fatigue level. Further, for example, in order to confirm the distraction level or to examine what the vehicle driver is currently concentrating on, the direction of the driver's line of sight can be tracked. For this purpose, it is necessary to accurately and reliably monitor the eyes of the vehicle driver. In order to be able to clearly recognize the eyes even in situations with poor visibility such as at night, an infrared camera is often used, particularly in combination with active infrared illumination. Further, a filter such as a band-pass filter that allows only infrared light to pass through can also be provided. Such a filter can be incorporated into the optical system of a corresponding camera or can be implemented digitally. Thereby, for example, an unpleasant glare effect caused by ambient light can be reduced or filtered, and only the relevant infrared light wavelength region can be passed through.

[0004] However, when using an infrared light source, reflection of the infrared spectrum may occur, for example, from eyeglasses worn by the vehicle driver. Infrared light from the surrounding environment may also enter the vehicle interior, potentially causing a glare effect. This can also result in projection. In such cases, the relevant camera image area may be overexposed or underexposed in the infrared spectrum, which could impair more reliable detection by the vehicle driver's eye.

[0005] From German Patent Application Publication No. 102017205386, a driver observation device and method for observing a driver in a vehicle in order to determine at least one position of the driver in the vehicle are known. The driver observation device is designed to illuminate the driver from two sides. The driver is illuminated from both the rear and the front, with infrared light being used in particular. A suitable light source can also project a light pattern structured like a grid onto the driver. This allows for a more reliable determination of the driver's head position using active triangulation. In this case, the camera used to detect the driver's face can be used simultaneously to detect the direction of line of sight.

[0006] Furthermore, German Patent Application Publication No. 102017216680 provides information on vehicle cockpit elements and automobiles equipped with such vehicle cockpit elements. This document describes an infrared light source designed as a surface-emitting means for illuminating the face of a vehicle driver. The surface-emitting means can be formed, for example, from an OLED film and incorporated into the steering wheel of a vehicle. The flat form of the infrared light source has the advantage of reducing or avoiding shadows on the driver's face due to diffuse light emission compared to a point light source. This allows for more reliable detection of the driver's face. However, this does not prevent reflection. Moreover, since the installation space within a vehicle to accommodate a suitable vehicle cockpit element is limited, considerable effort is required to incorporate the vehicle cockpit element into the corresponding vehicle.

[0007] International Publication No. 2019 / 215286 further discloses an apparatus and method for operating an object recognition system in the interior of an automobile, as well as an automobile. This apparatus uses an infrared light source incorporated into the automobile's display device to illuminate the face area of ​​a vehicle occupant with infrared light. The infrared light source comprises a plurality of IR light-emitting elements arranged in a matrix. The IR light-emitting elements can be controlled based on an evaluation of a camera image showing the illuminated face area.

[0008] German Patent No. 102014009143 further discloses an automotive camera device with controllable active illumination. An optical filter having at least two different filter regions with at least temporary differences in transparency is placed in the optical path of the light source of the active illumination. The transparency can be precisely increased or decreased to avoid light reflection in the camera image produced by the camera device.

[0009] Furthermore, in-cabin optical monitoring is known from U.S. Patent No. 9,290,146. This can be used to activate a vehicle's airbags only when a person is actually sitting in the seat assigned to each airbag. [Overview of the project] [Problems that the invention aims to solve]

[0010] The present invention aims to provide an improved vehicle occupant detection device that enables more reliable detection by the driver's eye. In this case, the above-mentioned vehicle occupant detection device needs to be easily integrated into the vehicle.

[0011] According to the present invention, this objective is achieved by a vehicle occupant detection device having the features of claim 1. Advantageous embodiments and developments, as well as vehicles equipped with such vehicle occupant detection devices, will become apparent from the dependent claims.

[0012] Comprehensively, the vehicle occupant detection device comprises at least one infrared light source, a camera, and a computing unit for controlling the infrared light source and evaluating the camera image generated by the camera, wherein the infrared light source and the camera are directed towards an observation area, the infrared light source is configured to illuminate at least one facial region of a person's face in the observation area, the camera is configured to detect at least the illuminated facial region, the computing unit is configured to recognize the direction of a person's gaze in the camera image, the infrared light source is formed by a plurality of infrared light-emitting elements incorporated into a display device and arranged in a matrix, the computing unit is further configured to activate IR light-emitting elements selected to emit infrared light, and the computing unit determines which IR light-emitting elements are selected to activate according to the evaluation result of the camera image. The present invention is improved in that the computing unit is further configured to control the matrix of IR light-emitting elements to emit a structured light pattern and to determine the relative position of a person's pupil to at least a portion of the light pattern in order to determine the direction of a person's gaze in the camera image.

[0013] By using the vehicle occupant detection device according to the present invention, the observed person, particularly the eye area, can be detected more reliably. The infrared light source includes a plurality of individually controllable IR light-emitting elements. These are arranged in a matrix, so that each IR light-emitting element primarily emits infrared light to a specific spatial area. Therefore, by selectively switching the IR light-emitting elements on and off, infrared light can be precisely projected to different spatial areas. This is used to brighten areas that should be illuminated by infrared light more brightly, and conversely, to prevent areas that should not be illuminated from being illuminated. This makes it possible to illuminate the scene more uniformly with infrared light, thereby producing camera images with fewer overexposed or underexposed areas.

[0014] To determine how to activate the IR light-emitting elements, a camera image of the scene is first captured. During the capture of this camera image, the infrared light source can be deactivated or activated. If the infrared light source is deactivated, the influence of infrared light emitted from the surrounding environment on the captured scene can be determined. Conversely, if the infrared light source is activated, for example, all IR light-emitting elements in the matrix may be activated, in which case the influence of the infrared light source itself on the illumination of the scene can be determined. Multiple camera images can also be evaluated sequentially, in which case each individual combination of IR light-emitting elements is activated when different camera images are captured. For example, if an object that reflects infrared light, such as glasses worn by a person, is within the observation area, a correspondingly overexposed area will appear in the camera image. Then, the computing unit controls the infrared light source and deactivates the IR light-emitting elements to eliminate the reflection. Furthermore, IR light-emitting elements can be activated to direct more infrared light to areas of the image that are too dark.

[0015] For example, the computing unit evaluates the captured camera images based on specific predetermined criteria. For instance, various performance indicators, also known as Key Performance Indicators (KPIs), can be predetermined, and the camera images can be evaluated based on these indicators. For this purpose, for example, the signal-to-noise ratio, the saturation of specific color channels, etc., can be evaluated for the entire camera image or for different areas of the camera image. Depending on which performance indicators are manifested in which image areas, the lighting settings are changed via the controller, thereby activating other IR light-emitting elements.

[0016] In addition, the vehicle occupant detection device according to the present invention can be easily incorporated into a vehicle. The matrix of IR light-emitting elements is a component of a display device, and the display device is an existing element that is often installed in the cockpits of modern vehicles, for example, to display information output via an instrument cluster. Therefore, the matrix of IR light-emitting elements cannot be directly seen by the observer, improving the appearance of the vehicle occupant detection device. This makes it possible to incorporate at least the infrared light source in an aesthetically pleasing manner and out of the viewer's sight.

[0017] IR light-emitting elements can be designed in various ways. For example, IR light-emitting elements can be made into point-like elements such as pixels or subpixels. At least some of the IR light-emitting elements can also be designed as flat or planar light-emitting elements, such as tape lights or square lights. Therefore, multiple such planar or tape lights can be arranged side by side to form the matrix described above. Point-like and flat light-emitting elements can also be combined.

[0018] During camera image capture, the IR light-emitting element can be operated continuously or in pulsed mode. In pulsed mode, the camera and infrared light source are synchronized so that infrared light is emitted precisely at the moment the camera image is generated. This allows for more reliable image capture even in darkness.

[0019] In this case, the infrared light output emitted to illuminate the scene, or observation area, remains nearly constant compared to using a single point light source. However, because the entire output is distributed over a relatively large area, the infrared light output per unit area decreases, reducing the amount of light reaching people and improving eye safety. This also has the effect of providing a more uniform heat distribution.

[0020] As described above, the computing unit controls the matrix of IR light-emitting elements to emit a structured light pattern, and is further configured to determine the relative position of the person's pupil to at least a portion of the light pattern in order to determine the direction of the person's gaze in the camera image. This makes it possible to determine the direction of the person's gaze more accurately. The light pattern generated by the infrared light source is projected onto the cornea of ​​the person being observed. Therefore, compared to a point light source, there are more IR structures on the cornea, allowing for a more accurate determination of the pupil's position relative to the light pattern. This allows for a more accurate determination of the relative position, and thus the direction of the gaze.

[0021] Conventional methods for detecting line of sight measure the distance from the pupil to the corneal reflection on the cornea. The location of the corneal reflection on the cornea is independent of the line of sight. However, if the observed person moves, the angle of incidence of infrared light to the eye changes, which can also displace the corneal reflection. In contrast, by using a structured light pattern instead of a point light source, the movement of the observed person's head can be tracked, thereby filtering out its effect on the line of sight. For example, if the observed person moves towards an infrared light source, the structured light pattern will become larger. Conversely, if the observed person moves up, down, left, or right in a plane, different parts of the eyeball will be illuminated, causing some areas of the structured light pattern to become larger and others to become smaller. Depending on which areas of the structured light pattern change in size and the intensity of that change, the computational unit can detect changes in head position and take this into account for calculating the line of sight.

[0022] Any conceivable shape and geometry can be used as the light pattern, including straight lines, curves, diamond grids, cross grids, or dot patterns.

[0023] An advantageous development of vehicle occupant detection devices is that, in the case of a display device designed as an LCD display, the matrix of IR light-emitting elements is incorporated into the backlight of the display device or forms at least part of the backlight, and in the case of a display device designed as an OLED display, the matrix of IR light-emitting elements corresponds to the matrix of pixels, so that each IR light-emitting element forms one additional subpixel of a pixel. Therefore, an infrared light source designed as a matrix of IR light-emitting elements can be incorporated into display devices of various technologies. Since the crystal of an LCD display is usually transparent to infrared light, IR light-emitting elements can be incorporated into an LCD display particularly easily and cost-effectively by incorporating them into the backlight or forming at least part of the backlight. LCD displays can be designed as monochrome displays or color displays. To ensure that observers can perceive the content displayed on the LCD display even under poor lighting conditions, the display is illuminated by a backlight. Therefore, light-emitting elements that emit, for example, white light, remain components of the backlight. In that case, IR light-emitting elements can be incorporated into the LCD display in addition to the originally existing light-emitting elements.

[0024] Using an OLED display allows for a particularly energy-efficient configuration of the vehicle occupant detection device. Such an OLED display includes a large number of pixels arranged in a matrix. The pixels also include subpixels of organic light-emitting diodes of various colors, such as red, green, blue, and / or cyan, magenta, and yellow. According to the present invention, at least some of the pixels of the OLED display are combined with additional subpixels that emit infrared light. In this case, compared to the LCD display described above, the light-emitting elements are positioned relatively far forward in the line of sight of the OLED display, allowing for highly efficient and high-intensity projection of light into the surrounding environment. This eliminates the need to pass light through deep display layers that can weaken the light intensity.

[0025] Furthermore, since various display technologies are installed in the vehicle, it is possible to incorporate an infrared light source into any conceivable type of display device regardless of the display specifications.

[0026] According to another advantageous embodiment of the vehicle occupant detection device according to the present invention, a grid is stored in the calculation unit, the grid represents at least a part of the observation area subdivided into a plurality of grid elements, and an assignment between the grid elements and the IR light emitting elements is stored in the calculation unit. When the calculation unit detects an IR reflection in the camera image, it is further configured to deactivate the IR light emitting element assigned to at least one grid element where the IR reflection is located, or at least reduce the brightness of the IR light emitting element.

[0027] In this way, the calculation unit can particularly easily identify which IR light emitting element to deactivate or reduce its brightness in order to prevent the occurrence of reflections or at least weaken the reflections to a level where the relevant parts of the face of the person being observed can be more reliably detected.

[0028] In another advantageous embodiment of the vehicle occupant detection device according to the present invention, a grid is stored in the calculation unit, the grid represents at least a part of the observation area subdivided into a plurality of grid elements, and an assignment between the grid elements and the IR light emitting elements is stored in the calculation unit. When the calculation unit detects an IR shadow in the camera image, it is further configured to activate the IR light emitting element assigned to at least one grid element where the IR shadow is located, or increase its brightness. This not only suppresses or at least weakens the IR reflection, but also particularly easily prevents the occurrence of IR shadows. Therefore, it is even possible to brighten the IR shadow or completely remove it from the camera image.

[0029] The assignment of grid elements and IR light-emitting elements is fixed. The assignment may be learned, for example, based on calibration measurements performed once in advance. These calibration measurements may be performed within the scope of the manufacture or development of the vehicle occupant detection device. The grid-based control of this matrix of IR light-emitting elements (both in relation to the avoidance of IR reflections and in relation to the avoidance of IR shadows) can be exemplified by the adaptive high-beam assist of a vehicle equipped with matrix headlights. That is, in the corresponding high-beam assist, the individual pixels of the matrix headlights are switched off for the area in which a preceding vehicle or an oncoming vehicle has been detected. Thus, here, when an IR reflection or an IR shadow is detected in the spatial area illuminated by the corresponding IR light-emitting element, the IR light-emitting element is accurately deactivated or activated.

[0030] In another advantageous embodiment of the vehicle occupant detection device, the computing unit is further configured to control the matrix of IR light-emitting elements to emit a structured light pattern and to determine the relative position of at least one body part of a person in the observation area by evaluating the distortion of the structured light pattern present in the camera image.

[0031] Therefore, the vehicle occupant detection device according to the invention can specify the positions of the individual body parts of the observed person in the observation area based on active triangulation. Thus, the structured light pattern generated by the infrared light source is projected not only onto the eye part of a person, but also onto the remaining part of the face or other parts of the body such as the shoulders, arms, hands, upper body, etc. This enables the provision of further support functions.

[0032] Therefore, in another advantageous embodiment of the vehicle occupant detection device according to the present invention, the computing unit is further configured to recognize changes in the relative position of body parts by analyzing at least two consecutive camera images and to detect the execution of gestures by a person therefrom. This enables the computing unit to recognize gestures, which can also be interpreted as operational actions. For example, a person may raise their hand and rotate it in the air, which can be interpreted as the rotation of a rotary control knob. Corresponding up-and-down or side-to-side swiping motions can be interpreted as operations on a sliding control knob.

[0033] The vehicle according to the present invention is equipped with at least one of the aforementioned vehicle occupant detection devices. At least one vehicle occupant detection device is present to detect the person driving the vehicle. However, additional vehicle occupant detection devices can also be incorporated into the vehicle to detect further vehicle occupants, such as a passenger in the front passenger seat.

[0034] Vehicles can be designed as, for example, passenger cars, trucks, vans, buses, etc. Vehicles do not necessarily have to be road vehicles. Therefore, vehicles can also be railway vehicles, ships, or aircraft.

[0035] According to an advantageous embodiment of the vehicle according to the present invention, the display device is formed by an instrument cluster display, a central display, a passenger seat display, or a head unit. Therefore, the vehicle occupant detection device according to the present invention can be incorporated into a variety of vehicles particularly flexibly and easily. When the matrix of infrared light sources, i.e., IR light-emitting elements, is incorporated into the instrument cluster, the driver can be illuminated over a particularly wide area by being positioned in the center in front of the driver. When incorporated into the head unit, multiple vehicle occupants can be illuminated simultaneously, for example, the driver, the passenger in the front seat, and a person seated in the center of the rear seat.

[0036] In another advantageous embodiment of the vehicle, the vehicle comprises at least one airbag and a control function for deploying the airbag, wherein a computing unit is configured to provide the relative position of at least one body part as an input variable to the control function, and the control function is configured to deploy the airbag only if, when an airbag deployment signal is generated, at least one body part is located within a predetermined spatial area assigned to the airbag. This avoids false or unnecessary deployment of the airbag. The predetermined spatial area is formed by the spatial area in which the corresponding body part would collide with the inflated airbag in the event of an accident. If the body part is not located within this spatial area, the airbag deployment is also unnecessary.

[0037] The control function is, for example, a dedicated airbag control device, or an airbag control program that runs on another control device or a central onboard computer.

[0038] Other advantageous embodiments of the vehicle occupant detection device and vehicle according to the present invention will also be apparent from exemplary embodiments which will be described in detail below with reference to the figures. [Brief explanation of the drawing]

[0039] [Figure 1] This is a schematic side view of a person monitored by the vehicle occupant detection device according to the present invention. [Figure 2] This figure shows a schematic and simplified layer structure of an LCD display including a matrix of IR light-emitting elements. [Figure 3] This is a schematic diagram of a pixel in a display device that includes subpixels configured to emit infrared light. [Figure 4] This is a schematic diagram showing a portion of the camera image of the detected person, the grid, and the assignment of IR light-emitting elements to the grid elements. [Modes for carrying out the invention]

[0040] Figure 1 shows a side view of a vehicle 15 according to the present invention. A person 5, in this case the vehicle driver, is present inside the vehicle. Person 5 is detected by a vehicle occupant detection device according to the present invention. This device comprises an infrared light source 1, a camera 2, and a computing unit 3. The vehicle occupant detection device tracks the line of sight of person 5, and this line of sight is used to provide driver assistance functions. Person 5 is illuminated with infrared light from the infrared light source 1 so that it can be better recognized even in poor lighting conditions. Therefore, the camera 2 is set to detect light in the infrared spectrum. For this purpose, the camera 2 may also have various filters.

[0041] In the exemplary embodiment shown in Figure 1, the infrared light source 1 is integrated into the display device 6, in this case the instrument cluster of the vehicle 15. Furthermore, the infrared light source 1 is formed by a matrix of IR light-emitting elements 7, as shown in Figure 2. The infrared light source 1 and camera 2 are directed towards the observation area 4 where the head of a person 5, and therefore the eyes to be observed, are located during the use of the vehicle 15.

[0042] Camera 2 can be positioned anywhere in the vehicle 15, as long as it can detect at least one facial region, preferably the eyes, of a person 5 within the observation area 4. For example, Camera 2 can be on the instrument panel of the vehicle 15, particularly in the area of ​​the display device 6, or similarly integrated into the display device 6 as shown in the figure.

[0043] The vehicle occupant detection device according to the present invention makes it possible to more reliably detect a person 5, especially their eyes and therefore their line of sight. During operation, IR reflections 13, as shown in Figure 4, may occur in the camera image generated by the camera 2, or individual image regions may have infrared shadows. In that case, the corresponding face region cannot be correctly recognized by the camera 2 or the computing unit 3. This can be prevented or at least attenuated by designing the infrared light source 1 as a matrix of IR light-emitting elements 7, as shown in Figure 2. In this way, individual IR light-emitting elements 7 are accurately activated or deactivated, and the formation of IR reflections 13 and / or IR shadows is prevented.

[0044] Figure 2 shows the configuration of the infrared light source 1 as part of the display device 6. In Figure 2, the display device 6 is designed as an LCD display. The display device 6 comprises, for example, four layers 16.1, 16.2, 16.3, and 16.4. The first layer 16.1 is the backlight 8 of the display device 6. For example, the second layer 16.2 is a thin-film transistor layer, also called a thin-film transistor or simply TFT. The third layer 16.3 is a layer made of liquid crystal, i.e., actual liquid crystal. The fourth layer 16.4 is a color filter layer or "pixel layer" containing a plurality of pixels 9 arranged in a matrix. The display device 6 may have further components not defined and illustrated in detail, such as polarizing filters, substrates, electrodes, cover layers, diffusion layers, or scattering layers.

[0045] The illustrated structure is illustrative, and the LCD display may have alternative structures known from the prior art, for example.

[0046] For example, in Figure 2, pixel 9 is shown to be larger in order to make subpixel 10 easier to recognize. Particularly preferably, an IR light-emitting element 7 is assigned to each pixel 9.

[0047] According to the present invention, individual IR light-emitting elements 7 can be designed as point light sources or flat light-emitting means. Figure 2 shows various possible embodiments in different areas (upper left, upper right, and bottom) of the backlight 8. The IR light-emitting elements 7 are indicated by hatching. The backlight 8 is further equipped with conventional light-emitting elements 17, such as cool white LEDs, LED strips, or similar, to improve the visibility of the display content of the display device 6 even under poor lighting conditions.

[0048] Pixel 9 contains various subpixels 10. For example, each pixel 9 contains subpixels 10 of red, green, and blue, indicated by the letters R, G, and B. Other color combinations are also possible, such as cyan, magenta, and yellow, indicated by the letters C, M, and Y. In the exemplary embodiment shown in Figure 2, at least one subpixel 10 of at least some pixels, or essentially all pixels 9, are transparent to infrared light, so that the infrared light emitted by the IR light-emitting element 7 can penetrate to the front portion of the display device 6 and thereby be projected onto a person 5. A void can also be provided to allow infrared light to pass through.

[0049] As shown in Figure 3, it is also possible to provide a separate subpixel 10 for transmitting infrared light. The corresponding subpixel 10 can be recognized by the character IR.

[0050] The exemplary embodiment shown in Figure 3 could also be a pixel 9 of a display device 6 designed as an OLED display. As shown in Figure 3, all possible arrangements and geometries of subpixels 10 are applicable. In that case, the example shown in Figure 3 is not limiting. Other arrangements and forms not shown in detail are also applicable.

[0051] Figure 4 shows the relationship between the portion of the face illuminated by the infrared light source 1 and the arrangement of the corresponding IR light-emitting elements 7 on or within the display device 6. Figure 4 shows the portion of the eyes of person 5 detected by camera 2. A grid 11 composed of multiple grid elements 12 is shown, and reference symbols are attached to only some of them for clarity. The computing unit 3 stores the assignment of which of these grid elements 12 is mainly illuminated by which IR light-emitting elements 7. When the computing unit 3 recognizes the presence of an IR reflection 13 by evaluating the corresponding camera image generated by camera 2, it determines which grid element 12 the IR reflection is located on. Next, the IR light-emitting elements 7 assigned to each grid element 12, shown as dark hatching in Figure 4, are deactivated. As a result, the IR reflection 13 disappears. Each IR light-emitting element 7 does not necessarily need to be deactivated; simply reducing its brightness is sufficient. This improves the recognition of person 5's pupil 14, allowing the IR camera to more reliably determine person 5's line of sight.

[0052] Additionally or alternatively, IR shadows can be avoided or at least reduced in a corresponding manner. [Prior art documents] [Patent Documents]

[0053] [Patent Document 1] German Patent Application Publication No. 102017205386 [Patent Document 2] German Patent Application Publication No. 102017216680 [Patent Document 3] International Publication No. 2019 / 215286 [Patent Document 4] German Patent Publication No. 102014009143 [Patent Document 5] U.S. Patent Publication No. 9290146

Claims

1. The system comprises at least one infrared light source (1), a camera (2), and a computing unit (3) for controlling the infrared light source (1) and evaluating the camera image generated by the camera (2), wherein the infrared light source (1) and the camera (2) are directed toward an observation area (4), the infrared light source (1) is configured to illuminate at least one facial region of a person (5) in the observation area (4), the camera (2) is configured to detect at least the illuminated facial region, and the computing unit (3) is configured to recognize the gaze direction of the person (5) in the camera image. The infrared light source (1) is incorporated into the display device (6) and is formed by a plurality of IR light-emitting elements (7) arranged in a matrix, the computing unit (3) is further configured to activate the IR light-emitting elements (7) selected to emit infrared light, and the computing unit (3) determines which IR light-emitting elements (7) to be activated according to the evaluation result of the camera image. The computing unit (3) controls the matrix of the IR light-emitting element (7) to emit a structured light pattern, and is further configured to determine the relative position of the pupil (14) of the person (5) and at least a portion of the light pattern in order to determine the direction of the person's (5) line of sight in the camera image. A vehicle occupant detection device characterized by the following features.

2. In the case of a display device (6) designed as an LCD display, the matrix of the IR light-emitting elements (7) is incorporated into the backlight (8) of the display device (6) or forms at least a part of the backlight; in the case of a display device (6) designed as an OLED display, the matrix of the IR light-emitting elements (7) corresponds to the matrix of pixels (9), so that each IR light-emitting element (7) forms an additional subpixel (10) of one pixel (9). The vehicle occupant detection device according to claim 1, characterized in that...

3. The calculation unit (3) stores a grid (11), the grid (11) representing a subdivision of at least a portion of the observation area (4) into a plurality of grid elements (12), the calculation unit (3) stores the assignment of grid elements (12) to IR light-emitting elements (7), and the calculation unit (3) is further configured to deactivate the IR light-emitting element (7) assigned to at least one grid element (12) where the IR reflection (13) is located, or to reduce the brightness of at least the IR light-emitting element (7), when an IR reflection (13) is detected in the camera image. A vehicle occupant detection device according to claim 1 or 2, characterized in that...

4. The calculation unit (3) stores a grid (11), the grid (11) representing a subdivision of at least a portion of the observation area (4) into a plurality of grid elements (12), the calculation unit (3) stores the assignment of grid elements (12) to IR light-emitting elements (7), and the calculation unit (3) is further configured to activate the IR light-emitting element (7) assigned to at least one grid element (12) where the IR shadow is located, or to increase the brightness of the IR light-emitting element (7), when an IR shadow is detected in the camera image. A vehicle occupant detection device according to any one of claims 1 to 3, characterized in that

5. The computing unit (3) is further configured to control the matrix of the IR light-emitting element (7) to emit a structured light pattern and to determine the relative position of at least one body part of the person (5) in the observation area (4) by evaluating the distortion of the structured light pattern present in the camera image. A vehicle occupant detection device according to any one of claims 1 to 4, characterized in that

6. The calculation unit (3) is further configured to recognize changes in the relative position of the body parts by analyzing at least two consecutive camera images, and to detect the execution of gestures by the person (5) from there. The vehicle occupant detection device according to claim 5, characterized in that

7. At least one vehicle occupant detection device according to any one of claims 1 to 6 A vehicle (15) characterized by the following.

8. The display device (6) is formed by an instrument cluster display, a central display, a passenger seat display, or a head unit. The vehicle (15) according to claim 7, characterized in that it is the vehicle according to claim 7.

9. The system is characterized by at least one airbag and a control function for activating the airbag, wherein the computing unit (3) is configured to provide the relative position of the at least one body part as an input variable for the control function, and the control function is configured to activate the airbag only when an airbag activation signal is generated and at least one body part is located within a predetermined spatial area assigned to the airbag. The vehicle (15) according to claim 7 or 8.

Images