Method for detecting an object and motor vehicle

Abstract

In order to detect objects on the basis of camera images of cameras (12, 14) having overlapping image regions, it should be possible to associate pixel groups in the individual images, for which a system of lines (L1, L2, L3,..., L48) fixed relative to the cameras is used, which emanate from the camera carriers. This is of particular interest when the motor vehicle (1) is used as a camera carrier for autonomous driving and for assisting the driver.

Description

Technical Field

[0001] This invention relates to a method for detecting an object from a carrier moving on the ground. The carrier can be understood, in particular, as a motor vehicle traveling on a street or other ground surface. The method for detecting the object can be part of a method for an autonomous vehicle, in which, upon detecting an object, control commands considering the detection are issued to the actuator units of the motor vehicle. It can also be part of a method for assisting a motor vehicle driver, wherein, similarly, commands are issued to the actuators of the motor vehicle based on the detection of the object; in this case, the actuators can be, in particular, output devices for warning and / or instructing the motor vehicle driver to react appropriately. The invention also relates to a motor vehicle with a corresponding configuration. Background Technology

[0002] An image sensor with multiple pixels for acquiring images is arranged / mounted / constructed on a carrier. This digital image sensor preferably captures pixel images in such a way that the pixels are arranged in rows and columns, generally in a checkerboard grid pattern.

[0003] In the case where a motor vehicle is used as the carrier moving on the ground, typically five cameras are specified, four of which are oriented in directions 90° apart from each other (e.g., forward, left, backward, and right). (The 90° value refers to the central axis of each camera; the cameras are preferably configured as optical cameras, allowing for a small deviation of up to 5° (therefore, this value includes an angular range of 85° to 95°)). Each of these cameras preferably has an image area / field of view of at least 150°, more preferably at least 160°, further preferably at least 170°, and particularly preferably at least 180°.

[0004] The fifth camera is also preferably forward-oriented, but located at a different height within the vehicle: the other four cameras can be positioned at approximately the height of the passenger seat or lower, while the fifth camera can be located in the top area of ​​the vehicle, capturing an additional forward area accordingly. Typically, the cameras are arranged in such a way that there is significant overlap / intersection between the image areas of the individual cameras.

[0005] The current problem is how to compare / verify the pixel images acquired by each image sensor. Typically, the image regions of an image sensor pair overlap (this overlap is particularly noticeable when a moving vehicle is stationary; therefore, it is a spatial overlap). Therefore, objects within the overlapping image regions should be reliably detectable from the two images of the image sensor pair based on the pixel data (color values, grayscale values, etc.) contained in their respective pixel images.

[0006] US 2014 / 0055573 A1 discloses a method of projecting 3D data captured by multiple cameras on a vehicle onto a plane, followed by calibration to overlay images from two cameras into a single image and capture 3D objects.

[0007] WO 2016 / 198059 A1 discloses the generation of virtual images of a vehicle environment using a camera surround-view system (camera environment detection system). Here, sampling points are defined based on a three-dimensional mesh, preferably defined in the space surrounding the vehicle, i.e., not directed at the vehicle itself. Such a mesh arrangement increases computational complexity. Therefore, it is desirable to detect objects as in real-time as possible, particularly based on images acquired substantially simultaneously by image sensors. This avoids the drawbacks of methods such as those according to US11 544 895 B2, in which a forward-looking camera first acquires an image of the environment, and then a backward-looking camera acquires the same image of the environment after passing the location. Summary of the Invention

[0008] Therefore, the objective of this invention is to simplify the inter-pixel correlation of different pixel images from multiple image sensors so as to identify objects as well and clearly as possible when image sensor pairs have overlapping image regions.

[0009] To accomplish this task, the present invention proposes a method for detecting an object having the features described in claim 1, comprising the following steps: - Each of the image sensors in at least one image sensor pair with overlapping image regions acquires at least one image; - Associate / map the pixels of the image from the first image sensor in an image sensor pair with overlapping image regions to the pixels of the image from the second image sensor in the same image sensor pair (thereby indirectly associating selected image structures, such as object edges, with corresponding image structures in another image via pixels). - Based on the associated pixel detection object, wherein in this invention, groups of pixels in an image are associated based on a line / curve search pattern preferably originating from the carrier in a top view, and the association is performed along said lines. These lines collectively constitute the search pattern. This search pattern is placed / projected into the image in a certain way, and then groups of pixels affected by the lines of the search pattern can be associated with the image structure in the overlapping area, thereby associating with objects presented in both images. Here, the concept of lines includes any type of direction, i.e., straight lines and curved lines, and may include both simultaneously in a search pattern.

[0010] The detection steps can also be described as follows: According to the pixel detection objects associated in this way, the pixel groups of the images of the first image sensor and the second image sensor are respectively mapped to lines, preferably emanating from the carrier in the top view, based on these lines, the pixels of the image of the first image sensor in the image sensor pair with overlapping image areas are associated with the pixels of the image of the second image sensor in the image sensor pair.

[0011] Solving this task also includes providing a motor vehicle with a control unit that performs the method.

[0012] This method consciously avoids defining a grid. Furthermore, pixel groups are now associated with a carrier (i.e., a moving vehicle in the example case), and they preferably originate from that carrier. Since image sensors are typically fixed to a carrier / vehicle, the spatial relationships between image regions of the image sensors are also fixed, facilitating inter-pixel correlation. For example, in an image from one image sensor, an imaginary line along a particular pixel can always correspond to the same imaginary line along that pixel in the image region of an adjacent image sensor (the second image sensor in an image sensor pair), capturing a specific region of the environment. Detected objects can include detecting any type of structure, surface, or terrain information. Within the scope of this invention, the image is simplified to a search pattern or simply a single line, thereby reducing the amount of data used for object recognition, and object recognition is performed based on this simplified image.

[0013] The primary application is in autonomous vehicles or when drivers are assisted by vehicle assistance systems, where the detected object is an obstacle. The obstacle preferably originates from the ground on which the vehicle moves. Obstacles can extend upwards (e.g., other vehicles, guardrails, construction site fencing) or downwards (e.g., potholes). The image sensor mounted on the vehicle is preferably located at a certain height above the ground. Light rays from the obstacle to the image sensor travel in a straight line, meaning that the foremost obstacle determines the pixel's data value (color value, grayscale value, etc.). Since the line originates from the vehicle, not necessarily from the image sensor, the pixel group is mapped to the line in the image. Now, if one mentally travels along such a pixel group, i.e., along the line, in the image captured by the image sensor, the obstacle is typically the first object with predetermined attributes. These predetermined attributes may include a significant difference in grayscale value between the pixel and the surrounding environment (e.g., a curb is light gray, while paved asphalt is dark gray); and adjacent pixels on the line used to identify objects can also be used to detect whether they are the target object (obstacle).

[0014] According to a first preferred embodiment of the invention, in the top view, all lines originating from the carrier are radial rays of straight lines originating from a center point. The center point can be located within the carrier or as an imaginary point below the carrier (e.g., on the road surface below the vehicle), but this point moves with the carrier. In the top view, the lines still originate from the carrier. These radial rays are particularly easy to calculate and can have predetermined angular intervals (e.g., from 0.2° to 2°, preferably approximately between 0.3° and 0.7°, e.g., 0.5°). However, the radial rays do not necessarily have uniform angular intervals. Using such straight radial rays, pixels in images acquired by adjacent image sensors (as a pair of image sensors with overlapping image areas) can be correlated particularly easily.

[0015] Alternatively, at least a portion of the lines can be specified to have curvature. This allows for the detection of objects that are occluded in the image of one image sensor but visible to adjacent image sensors, thus enabling "angle-based viewing" from the perspective of the first image sensor.

[0016] According to an alternative approach, additional pixels are defined by branches of lines originating from the carrier in the top view, with the branches positioned outside the carrier. This takes into account the fact that the distance between lines increases with distance from the carrier / vehicle, thus the branches ensure sufficient density of object capture even at greater distances from the carrier / vehicle. Branching can begin from a predetermined distance / within a predetermined distance range; further branches may occur beyond one or two existing branches.

[0017] According to a preferred embodiment, when acquiring an image, the acquisition time is stored as a so-called timestamp. The timestamp is taken into account when associating pixels. This embodiment takes into account the fact that, if the overlapping image area is large enough, the images in an image sensor pair will still overlap even if the sensors do not acquire their images perfectly simultaneously. This situation occurs when one image is acquired when the carrier / vehicle is in a first position, and the second image is acquired when the carrier / vehicle moves to a second position (the time interval may be, for example, 50 milliseconds to 200 milliseconds). However, since the lines are associated with the carrier / vehicle, i.e., defined in the carrier's world coordinate system, the world coordinate system in the second case (subsequent movement) can be reversed to the world coordinate system in the first case (when the first image was acquired), or vice versa. This makes association possible even when the image sensors are not perfectly synchronized.

[0018] This embodiment demonstrates a significant advantage of the lines defined by the present invention, originating from the carrier / motor vehicle (in a top view). Preferably, in addition to timestamps, the carrier's speed is also taken into account, or alternatively or additionally, the wheel steering angle (such as the steering angle measured in a motor vehicle) and / or the carrier's yaw angle (measured by corresponding sensors), thereby achieving particularly high correlation accuracy.

[0019] According to a preferred embodiment of the invention, in addition to an image sensor, radar and / or lidar devices (Light detection and ranging or Light imaging, detection and ranging, roughly translated as "light-based object recognition and ranging") are used, and their measurement results are utilized when associating image pixels and / or detecting objects. Specifically, a three-dimensional space is also constructed and filled with virtual lines / search patterns. This also ensures high-precision and reliable object detection.

[0020] The motor vehicle of the present invention includes at least two digital sensors for acquiring data within an acquisition range, at least one of which is an image sensor, preferably both being image sensors for acquiring pixel images, the at least two (image) sensors acting as a sensor pair having overlapping acquisition ranges / image regions in the simultaneously acquired data / images. The motor vehicle includes an evaluation device for evaluating the data / images, the evaluation device being configured to: define at least one two-dimensional, preferably three-dimensional, virtual acquisition space / image space by groups of pixels in the image / data, each associated with a line (preferably emanating from the motor vehicle in a top view); and based on the groups of pixels in the image of the (image) sensor pair having overlapping acquisition ranges / image regions (thus associated with the line), associate pixels of the image of the first image sensor in the (image) sensor pair with overlapping acquisition ranges / image regions with pixels of the image / data of the second sensor in the (image) sensor pair; the evaluation device is also configured to detect objects based on the pixels thus associated.

[0021] The method of the present invention is performed on a motor vehicle, and therefore the above-mentioned advantages also apply to it.

[0022] According to an advantageous embodiment, the motor vehicle is configured for autonomous driving, and its evaluation equipment is configured to at least identify obstacles as objects.

[0023] According to another advantageous embodiment, the motor vehicle includes an optical camera as an image sensor in other known ways, preferably comprising at least four cameras, each having an image area of ​​at least 150°, and more preferably, a first camera oriented forward, and a fifth camera, also oriented forward, but mounted at a different vehicle height than the first camera.

[0024] The present invention can also be implemented as a motor vehicle having at least two digital sensors for acquiring images within their respective acquisition ranges, wherein at least one sensor is a radar and / or lidar device, and the at least two sensors, as a sensor pair, have overlapping acquisition ranges (spatial overlap) in the simultaneously acquired images, and are provided with an evaluation device for evaluating the images, the evaluation device being configured to: define at least one two-dimensional, preferably three-dimensional, virtual acquisition space by a set of points in the images associated with lines preferably originating from the motor vehicle in a top view; and associate points of the first sensor image of the sensor pair with points of the sensor pair having overlapping acquisition ranges (in this way or in association with lines) with points of the second sensor image of the sensor pair; the evaluation device is also configured to detect objects based on the points thus associated.

[0025] For application scenarios or situations where this method may occur but are not explicitly described here, it may be specified to output error messages and / or requests for user feedback according to this method, and / or set default settings and / or predetermined initial states.

[0026] The present invention also includes a control device for a motor vehicle. This control device may have a data processing apparatus or a processor device configured to perform an embodiment of the method of the present invention. For this purpose, the processor device may include at least one microprocessor and / or at least one microcontroller and / or at least one application-specific integrated circuit (ASIC) and / or at least one field-programmable gate array (FPGA) and / or at least one digital signal processor (DSP). As a microprocessor, a central processing unit (CPU), a graphics processing unit (GPU), or a neural processing unit (NPU) may be used, in particular. Furthermore, the processor device may have program code configured to perform an embodiment of the method of the present invention when executed by the processor device. This program code may be stored in the data memory of the processor device. The processor device may, for example, be based on at least one circuit board and / or at least one system-on-a-chip (SoC).

[0027] The present invention also includes improvements to the motor vehicle of the present invention, which have features related to improvements to the method of the present invention. Therefore, corresponding improvements to the motor vehicle of the present invention will not be described again here.

[0028] The motor vehicle of the present invention is preferably configured as an automobile, particularly a passenger car or commercial vehicle, or a bus or motorcycle.

[0029] As another solution, the present invention also includes a computer-readable storage medium containing program code that, when executed by a computer or computer cluster, causes it to perform an embodiment of the method of the present invention. The storage medium may be provided at least partially as non-volatile data storage (e.g., flash memory and / or solid-state drives) and / or at least partially as volatile data storage (e.g., random access memory). The storage medium may be arranged in a computer or computer cluster. However, the storage medium may also operate on the Internet, for example, as a so-called app store server and / or cloud server. A processor circuit with, for example, at least one microprocessor may be provided via the computer or computer cluster. The program code may be provided as binary code and / or assembly code and / or source code in a programming language (e.g., C) and / or program scripts (e.g., Python).

[0030] The invention also includes combinations of features of the described embodiments. Therefore, the invention also includes implementations having multiple combinations of features of the described embodiments, provided that these embodiments are not described as mutually exclusive. Attached Figure Description

[0031] Embodiments of the present invention are described below. Therefore: Figure 1A illustrates a motor vehicle according to an embodiment of the present invention, wherein a method according to an embodiment of the present invention may be implemented; Figure 1B shows the structure of the data processing equipment for the motor vehicle in Figure 1A; Figure 2 A top view of the motor vehicle in Figure 1A is shown, along with lines defined according to an embodiment of the present invention and the object to be detected; Figure 3A shows Figure 2 An example image captured by the first camera of the vehicle. Figure 3B shows Figure 2 Exemplary images captured by the second camera of the vehicle; Figure 4 This illustrates a first alternative direction of the lines in a first alternative embodiment of the method of the present invention; Figure 5 This illustrates an alternative direction of the lines in a second alternative embodiment of the method of the present invention. Detailed Implementation

[0032] The embodiments explained below are preferred embodiments of the present invention. In these embodiments, each of the described embodiment components represents an independent, separately conceivable feature of the invention, which can also independently further develop the invention. Therefore, this disclosure should also include combinations other than the illustrated combination of features. Furthermore, the embodiments can also be supplemented by other features already described in the invention.

[0033] In the figures, the same reference numerals indicate elements that have the same function.

[0034] The vehicle shown in Figure 1A, indicated by 1 throughout, includes a central data processing unit 10 and five cameras. The front-facing camera 12 captures the area in front of the vehicle, the second camera 14 (position not precisely marked) captures the right-side environment of the vehicle, the rear-facing third camera 16 captures the area behind the vehicle, the left-side camera 18 captures the left-side area of ​​the vehicle, and the fifth camera 20 also captures the area in front of the vehicle, but from a higher altitude than camera 12, thus obtaining additional information. Images from cameras 12, 14, 16, 18, and 20 are input to a data processing device 10, which, according to the configuration shown in FIG1B, includes: a receiving device 10-1 for receiving images from the cameras; an association device 10-2 for associating the images with each other according to association rules / mapping rules (which are based on lines explained below); a detection device 10-3 for detecting objects based on the association; and an actuator 10-4 for controlling the actuators of the motor vehicle 1 (a single actuator 20 is shown exemplarily in FIG1A) based on the detection of objects and / or the absence of detected objects or the absence of detected objects within a predetermined field of view.

[0035] Figure 2 The diagram shows a top view of vehicle 1 in Figure 1A, using a radial pattern fixed relative to the vehicle's coordinate system. Camera 12 captures an image region extending from the left edge 12-LR to the right edge 12-RR, and camera 14 captures an image region extending from the left imaging edge 14-LR to the right imaging edge 14-RR. These two image regions overlap.

[0036] Starting from a (virtual) center point M located beneath the vehicle, radial rays should capture / cover the ground in front of, to the sides of, and behind the vehicle. These rays are straight lines, correspondingly labeled as lines L1, L2, L3, L4, L5, L6, etc., up to line L48. The image area of ​​camera 12 at least partially includes lines L41 to L1 and L1 to L9. The image area of ​​camera 14 at least partially includes lines L3 to L22. Accordingly, the image areas of the first camera 12 and the second camera 14 overlap in the area of ​​lines L3 to L9. These radial rays together constitute the search pattern.

[0037] Figure 3A shows an exemplary image B12 (“photo”) captured by camera 12. Figure 3B shows, as an example, the corresponding image B14 captured by the second camera 14.

[0038] In image B12, each of the virtual lines L41 to L48 and L1 to L9 in the image area is now associated with a pixel group depicting that line. These pixel groups are here labeled as PG12-41 (corresponding to line L41), PG12-42 (corresponding to line L42), etc., and PG12-1 (corresponding to line L1), PG12-2 (corresponding to line L2), etc. Correspondingly, pixel groups can be seen in image B14, exemplarily labeled as pixel groups PG14-7, PG14-8, PG14-9, and PG14-10, which correspond to lines L7, L8, L9, and L10, respectively. Since cameras 12, 14, 16, 18, and 20 are fixed relative to motor vehicle 1 (body / vehicle structure), pixel group PG12-7 in the image acquired from camera 12 (as shown in Figure B12) corresponds to pixel group PG14-7 in the image acquired from camera 14 (as shown in Figure B14). The orientation of pixel groups PG12-1, PG12-2, etc. in image B12, and the orientation of pixel group PG14-7, etc. in image B14, are always the same relative to their respective image frames in all images; only the corresponding pixel data values ​​carried by the pixels differ.

[0039] Therefore, pixel groups PG12-7 and PG14-7 also correspond to line L7, making pixel groups PG12-7 and PG14-7 equivalent to each other. This allows the identification of object 2 in the image. Starting from the imaginary vehicle at the lower edge of images B12 and B14, one can travel along pixel group P12-7 and identify its corner point 12E. Similarly, one can travel along pixel group PG14-7 according to the arrow in Figure 3B and identify its corner point 2E. Ideally, the pixel data value at image point 2E in Figure 3A should be the same for cameras 12 and 14 of the same type, or for different cameras, it should be correlated according to known association rules. Device 10-2 of the data processing device 10 of vehicle 1 correlates pixel groups PG12-7 and PG14-7, particularly the pixel data value at corner point 2E shown for object 2. Therefore, object 2 can now be identified in device 10-3 of data processing device 10, and the actuators in motor vehicle 1 can be controlled accordingly via device 10-4 of data processing device 10. For example, it is necessary to prevent motor vehicle 1 from turning right so as not to collide with object 2. Furthermore, based on the detection of corner point 2E, the subsequent leftward direction of object 2 can be detected, and motor vehicle 1 can be set to bypass object 2. Object 2 may simply represent a sidewalk, with its front edge and corner point 2E as its boundary (curb). Within the scope of this invention, the image is simplified to a search pattern or simply a single line, thereby reducing the amount of data, and object recognition is performed based on this simplified data.

[0040] Images B12 and B14 are assumed to have been acquired simultaneously. However, slight time differences can be considered, and the corresponding timestamps should be recorded when the images are acquired. Based on this information (preferably combined with further information about the speed, steering angle, and yaw angle of vehicle 1), correlation can be performed in device 10-2. In this case, the lines of the search pattern are adjusted by calculation to take into account vehicle motion, thereby ensuring that pixel groups (in this example, pixel groups PG12-7 and PG14-7) do indeed correspond to each other.

[0041] In addition to using lines L1, L2, and L3, according to Figure 4 It can be stipulated that lines K1, K2, K3, K4, etc., are defined starting from vehicle 1 in the top view. These lines are partly straight, such as lines K1, K8, and K9, and partly curved, so as to better capture certain areas ("around the corner").

[0042] Because according to Figure 2In a radial arrangement emanating from the center point M, the pixel density captured by the pixel group decreases; therefore, it can also be specified that lines L1, etc., branch in a direction away from the vehicle 1. This is in Figure 5 The text is presented as an example: Figure 2 A variant of line L1, variant L1', is shown, where line L1' continues as line L1'a, but also branches into line L1'b. Line L1'a continues as line L1'aa and branches into line L1'ab. Line L1'b further branches into lines L1'ba and L1'bb.

[0043] It is possible that only one line in the first branch branches again. It is also possible that there are more than one double branch (triple branch, quadruple branch, etc.).

[0044] The search pattern does not necessarily have to be centered within the body / supporting structure / vehicle shell area. The same or similar search patterns described herein can also be defined, which are fixed relative to the body / supporting structure / vehicle shell but located outside it.

[0045] Acquiring measurement data (related to acquiring images via an image sensor) may also include acquiring metadata (upper-level information) about the measurement points, and where available, intermediate information (such as covariance, classification information, object recognition information), or metadata about the sensor itself, such as the boundaries of the image region.

[0046] Overall, these examples demonstrate how to provide correlation between measurements from multiple cameras (or sensors in general) by using a fixed radial scanning pattern.

Claims

1. A method for detecting an object (2) from a carrier (1) moving on the ground, wherein a plurality of image sensors (12, 14, 16, 18, 20) are configured on the carrier to acquire pixel images, wherein at least one image sensor provides overlapping image regions in simultaneously acquired images, the method comprising the steps of: - Each of the image sensors in (12, 14) acquires at least one image by at least one image sensor pair with overlapping image regions; - Associate the pixels of the image (B12) of the first image sensor (12) in the image sensor pair with overlapping image regions with the pixels of the image (B14) of the second image sensor (14) in the same image sensor pair; - According to the associated pixel detection object, the pixel group (PG12-7; PG14-7) of the image is mapped to a line (L1, L2, L3, ..., L7, ..., L48) preferably originating from the carrier (1) in the top view, and the association is based on the line.

2. The method according to claim 1, characterized in that, The object is identified as an obstacle, which originates from the ground on which the carrier moves, and the obstacle is the first object (2) with a predetermined attribute on at least one line (L7).

3. The method according to claim 1 or 2, characterized in that, In the top view, all the lines originating from the carrier (1) are straight radial rays (L1, L2, L3, ..., L48) originating from the center point (M).

4. The method according to claim 1 or 2, characterized in that, At least some of the lines (K2, K3, K4, ..., K11, K12) have curvature.

5. The method according to claim 1 or 2, characterized in that, In the top view, the other pixels of the pixel group are defined by branching from the line (L1') originating from the carrier, with the branch located on the outside of the carrier.

6. The method according to any one of the preceding claims, characterized in that, When acquiring images, the acquisition time is stored as a timestamp and the timestamp is considered when associating pixels, wherein preferably the speed of the carrier (1) and / or the steering angle of the wheels of the carrier (1) and / or the yaw angle of the carrier (1) and / or the three-dimensional vehicle motion measurement of the inertial measurement unit are also considered.

7. The method according to any one of the preceding claims, characterized in that, In addition to image sensors, radar and / or lidar devices are used to utilize the measurement results of radar and / or lidar devices when correlating image pixels and / or detecting objects.

8. A motor vehicle (1) having at least two digital image sensors (12, 14, 16, 18, 20) for acquiring pixel images, wherein the at least two image sensors (12, 14) as an image sensor pair have overlapping image regions in simultaneously acquired images (B12, B14), the motor vehicle having an evaluation device (10) for evaluating the images, the evaluation device being configured to: define at least one two-dimensional, preferably three-dimensional, virtual image space by pixel groups of images, the pixel groups being associated with lines (L1, L2, L3; K1, K2; L1',...) preferably emanating from the motor vehicle (1) in a top view; and, based on the pixel groups of the images of the image sensor pair having overlapping image regions, associate pixels of the image of the first image sensor in the corresponding image sensor pair with pixels of the image of the second image sensor in the corresponding image sensor pair; wherein, The evaluation device (10) is also configured to detect objects (2) based on such associated pixels (PG12-7; PG14-7).

9. The motor vehicle (1) according to claim 8, characterized in that, The vehicle is configured for autonomous driving, and the evaluation device (10) is configured to at least identify obstacles (2) as objects.

10. The motor vehicle (1) according to claim 8 or 9, characterized in that, The vehicle has optical cameras (12, 14, 16, 18, 20) as image sensors. The optical cameras preferably include at least four cameras (12, 14, 16, 18), each with an image area of ​​at least 150°. More preferably, the first camera (12) is oriented forward, and a fifth camera (20) is also oriented forward, which is mounted at a different height on the vehicle (1) than the first camera (12).

11. A motor vehicle having at least two digital sensors (12, 14, 16, 18, 20) for acquiring images within their respective acquisition ranges, at least one sensor being a radar and / or lidar device, the at least two sensors (12, 14) having overlapping acquisition ranges in simultaneously acquired images (B12, B14) as a sensor pair, the motor vehicle having an evaluation device (10) for evaluating the images, the evaluation device being configured to: define at least one two-dimensional, preferably three-dimensional, virtual acquisition space by a set of points of the images, the set of points being associated with lines (L1, L2, L3; K1, K2; L1', ...) preferably emanating from the motor vehicle (1) in a top view; associate points of the image of the first sensor in the respective sensor pair with points of the image of the sensor pair having overlapping acquisition ranges; the evaluation device (10) is also configured to detect an object (2) based on the points thus associated.

Images