Collision detection method and device based on automatic parking obstacle map

Abstract

The invention relates to a collision detection method and device based on an automatic parking obstacle map, and the method comprises the steps: obtaining laser radar original point cloud data of a vehicle, and converting the laser radar original point cloud data into an obstacle binary matrix; a target collision detection strategy is determined according to the parking scene, corresponding vehicle parameters are obtained, the obstacle expansion radius of the target vehicle is calculated, and then corresponding structural elements are calculated; image expansion operation is carried out based on the barrier binary matrix and the structural data, a barrier map binary image matrix is generated, a barrier map is constructed, and the current position of the vehicle is acquired, so that a matrix coordinate point set of the vehicle on the barrier map under a rectangular frame sampling strategy or a multi-circle strategy is calculated according to the current position; and carrying out collision detection on the vehicle based on the matrix coordinate point set and a collision detection strategy. Therefore, the problems that in the prior art, part of passable areas can be lost in narrow and complex scenes, and the requirements for computing power and equipment are high when the occupied grid is used for describing a map are solved.

Description

Technical Field

[0001] The present application relates to the technical field of collision detection in an automatic parking system, and in particular to a collision detection method and device based on an automatic parking obstacle map. Background Art

[0002] With the rapid development of autonomous driving technology, the Automated Parking Assist (APA) system, as a key technology to enable autonomous vehicles to automatically park into parking spaces, can effectively improve parking efficiency and user experience, and has become a focus of research in academia and industry.

[0003] As a key technology of the APA system, path planning technology has a direct impact on the accuracy, safety and reliability of vehicle parking. In complex parking environments, how to make full use of the traversable area and improve the efficiency of automatic parking is the key to the research, so high requirements are placed on obstacle map generation and collision detection during parking path planning.

[0004] At present, conventional obstacle perception describes the vehicle's surrounding environment in the form of a bounding box. Due to the irregularity of obstacles, a part of the traversable area will be lost in narrow and complex scenes. Using occupancy grids to describe the map has high requirements on computing power and equipment, which needs to be solved urgently. Summary of the invention

[0005] The present application provides a collision detection method and device based on an automatic parking obstacle map to solve the problems that the prior art loses part of the traversable area in narrow and complex scenes, and that the use of an occupancy grid to describe the map has high requirements on computing power and equipment.

[0006] The first aspect of the present application provides a collision detection method based on an automatic parking obstacle map, comprising the following steps: obtaining laser radar raw point cloud data of a target vehicle, and converting the laser radar raw point cloud data into an obstacle binary matrix; determining a corresponding target collision detection strategy according to the parking scene of the target vehicle to obtain vehicle parameters corresponding to the target collision detection strategy, and calculating the obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy; performing an image expansion operation based on the obstacle binary matrix and the structural data to generate an obstacle map. Figure 2 value image matrix, and use the obstacle map Figure 2The target vehicle is constructed by using a value image matrix to construct a corresponding obstacle map, and the current position of the target vehicle is obtained to calculate a matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, and a collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy is used to perform collision detection on the target vehicle.

[0007] Optionally, in one embodiment of the present application, the acquiring of the target vehicle's laser radar raw point cloud data and converting the laser radar raw point cloud data into an obstacle binary matrix include: eliminating the ground point cloud in the three-dimensional laser radar of the laser radar raw point cloud data and the point cloud that does not meet the preset height and distance requirements to obtain a three-dimensional residual point cloud, and removing the height value of the three-dimensional residual point cloud to generate a corresponding first two-dimensional point cloud; eliminating the point cloud that does not meet the preset distance requirement in the two-dimensional laser radar of the laser radar raw point cloud data to obtain a second two-dimensional point cloud; translating the first two-dimensional point cloud and the second two-dimensional point cloud to obtain a translated point set; and constructing the obstacle binary matrix based on the translated point set and a preset pixel map step size.

[0008] Optionally, in one embodiment of the present application, the target collision detection strategy corresponding to the parking scene of the target vehicle is determined to obtain vehicle parameters corresponding to the target collision detection strategy, and the obstacle expansion radius of the target vehicle is calculated based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy, including: when the target collision detection strategy is the rectangular border sampling strategy, determining the sampling step length corresponding to the rectangular border sampling strategy, so as to perform the sampling step length and the vehicle parameters according to the sampling step length. Sampling operation is performed to obtain a corresponding vehicle bounding box point set; when the target collision detection strategy is the multi-circle strategy, the vehicle rectangle corresponding to the target vehicle is divided into multiple sub-rectangles according to the vehicle parameters and preset division requirements, and the circumscribed circle radius and the center coordinates of each sub-rectangle in the multiple sub-rectangles are calculated, so as to construct a two-dimensional point set of the center corresponding to the multi-circle strategy based on the circumscribed circle radius and the center coordinates; based on the preset safety distance, the vehicle bounding box point set and the two-dimensional point set of the center, the obstacle expansion radius corresponding to the rectangle bounding box sampling strategy or the multi-circle strategy is calculated, and the corresponding structural element is calculated according to the obstacle expansion radius.

[0009] Optionally, in one embodiment of the present application, the image dilation operation is performed based on the obstacle binary matrix and the structure data to generate an obstacle map. Figure 2 value image matrix, and use the obstacle map Figure 2The method comprises: using the structure element of the target collision detection strategy to perform an image expansion operation on the obstacle binary matrix to obtain the obstacle map corresponding to the target collision detection strategy; and obtaining the current position of the target vehicle to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position. Figure 2 value image matrix; based on the current position of the target vehicle and the preset pixel map step size, calculate the matrix coordinate point set of the vehicle border point set or the circle center two-dimensional point set on the obstacle binary image matrix.

[0010] Optionally, in one embodiment of the present application, the collision detection strategy based on the matrix coordinate point set and the target collision detection strategy performs collision detection on the target vehicle, including: when the target collision detection strategy is the rectangular border sampling strategy, judging whether all points in the matrix coordinate point set corresponding to the rectangular border sampling strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle border point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle border point set, then the target vehicle has not collided; when the target collision detection strategy is the multi-circle strategy, judging whether all points in the matrix coordinate point set corresponding to the multi-circle strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the circle center two-dimensional point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the circle center two-dimensional point set, then the target vehicle has not collided.

[0011] The second aspect of the present application provides a collision detection device based on an automatic parking obstacle map, including: a conversion module, used to obtain the laser radar original point cloud data of the target vehicle, and convert the laser radar original point cloud data into an obstacle binary matrix; a calculation module, used to determine the corresponding target collision detection strategy according to the parking scene of the target vehicle, so as to obtain the vehicle parameters corresponding to the target collision detection strategy, and calculate the obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy; a collision detection module, used to perform an image expansion operation based on the obstacle binary matrix and the structural data to generate an obstacle map. Figure 2 value image matrix, and use the obstacle map Figure 2The target vehicle is constructed by using a value image matrix to construct a corresponding obstacle map, and the current position of the target vehicle is obtained to calculate a matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, and a collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy is used to perform collision detection on the target vehicle.

[0012] Optionally, in one embodiment of the present application, the conversion module includes: a first elimination unit, used to eliminate the ground point cloud and the point cloud that does not meet the preset height and distance requirements in the three-dimensional laser radar of the laser radar original point cloud data to obtain the three-dimensional residual point cloud, and remove the height value of the three-dimensional residual point cloud to generate the corresponding first two-dimensional point cloud; a second elimination unit, used to eliminate the point cloud that does not meet the preset distance requirement in the two-dimensional laser radar of the laser radar original point cloud data to obtain the second two-dimensional point cloud; a translation unit, used to translate the first two-dimensional point cloud and the second two-dimensional point cloud to obtain a translation point set; a construction unit, used to construct the obstacle binary matrix based on the translation point set and the preset pixel map step size.

[0013] Optionally, in one embodiment of the present application, the calculation module includes: a sampling unit, which is used to determine the sampling step corresponding to the rectangular border sampling strategy when the target collision detection strategy is the rectangular border sampling strategy, so as to perform a sampling operation according to the sampling step and the vehicle parameters to obtain a corresponding vehicle border point set; a division unit, which is used to divide the vehicle rectangle corresponding to the target vehicle into multiple sub-rectangles according to the vehicle parameters and preset division requirements when the target collision detection strategy is the multi-circle strategy, and calculate the circumscribed circle radius and the center coordinates of each sub-rectangle in the multiple sub-rectangles, so as to construct a two-dimensional point set of the center corresponding to the multi-circle strategy based on the circumscribed circle radius and the center coordinates; an operation unit, which is used to calculate the obstacle expansion radius corresponding to the rectangular border sampling strategy or the multi-circle strategy based on a preset safety distance, the vehicle border point set and the two-dimensional point set of the center, and calculate the corresponding structural element according to the obstacle expansion radius.

[0014] Optionally, in one embodiment of the present application, the collision detection module includes: an image expansion unit, configured to perform an image expansion operation on the obstacle binary matrix using the structural element of the target collision detection strategy to obtain an obstacle map corresponding to the target collision detection strategy. Figure 2 A binary image matrix; a computing unit, used to calculate the matrix coordinate point set of the vehicle border point set or the circle center two-dimensional point set on the obstacle binary image matrix based on the current position of the target vehicle and a preset pixel map step.

[0015] Optionally, in one embodiment of the present application, the collision detection module further includes: a first judgment unit, used to determine, when the target collision detection strategy is the rectangular border sampling strategy, whether all points in the matrix coordinate point set corresponding to the rectangular border sampling strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle border point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle border point set, then the target vehicle has not collided; a second judgment unit, used to determine, when the target collision detection strategy is the multi-circle strategy, whether all points in the matrix coordinate point set corresponding to the multi-circle strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the circle center two-dimensional point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the circle center two-dimensional point set, then the target vehicle has not collided.

[0016] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the collision detection method based on the automatic parking obstacle map as described in the above embodiment.

[0017] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned collision detection method based on the automatic parking obstacle map.

[0018] The fifth aspect of the present application provides a computer program product, including a computer program, which is executed to implement the above-mentioned collision detection method based on the automatic parking obstacle map.

[0019] Therefore, the embodiments of the present application have the following beneficial effects:

[0020] The embodiments of the present application can obtain the original laser radar point cloud data of the target vehicle and convert the original laser radar point cloud data into an obstacle binary matrix; determine the corresponding target collision detection strategy according to the parking scene of the target vehicle to obtain the vehicle parameters corresponding to the target collision detection strategy, and calculate the obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular frame sampling strategy and a multi-circle strategy; perform an image expansion operation based on the obstacle binary matrix and the structural data to generate an obstacle map. Figure 2 Value image matrix, and use the obstacle map Figure 2The corresponding obstacle map is constructed based on the value image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular border sampling strategy or the multi-circle strategy according to the current position, and the target vehicle is subjected to collision detection based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy. The present application generates an obstacle map and performs efficient and rapid collision detection by accurately describing the obstacles and selecting the appropriate collision detection form according to the scene. Thus, the problems that the prior art loses part of the passable area in narrow and complex scenes, and the use of the occupied grid to describe the map have high requirements on computing power and equipment are solved.

[0021] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0023] Figure 1 A flowchart of a collision detection method based on an automatic parking obstacle map provided according to an embodiment of the present application;

[0024] Figure 2 A schematic diagram of a vehicle surrounded by a multi-circle method provided for an embodiment of the present application;

[0025] Figure 3 A schematic diagram of the execution logic of a collision detection method based on an automatic parking obstacle map provided for one embodiment of the present application;

[0026] Figure 4 A schematic diagram of rectangular frame sampling points provided for an embodiment of the present application;

[0027] Figure 5 A schematic diagram of an obstacle expansion provided for one embodiment of the present application;

[0028] Figure 6 A schematic diagram of collision detection of a rectangular frame sampling strategy provided for an embodiment of the present application;

[0029] Figure 7 is an example diagram of a collision detection device based on an automatic parking obstacle map according to an embodiment of the present application;

[0030] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0031] Among them, 10-collision detection device based on automatic parking obstacle map; 100-conversion module, 200-computing module, 300-collision detection module; 801-memory, 802-processor, 803-communication interface. DETAILED DESCRIPTION

[0032] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0033] The following describes the collision detection method and device based on the automatic parking obstacle map of the embodiment of the present application with reference to the accompanying drawings. In response to the problems mentioned in the above background technology, the present application provides a collision detection method based on the automatic parking obstacle map. In this method, the original laser radar point cloud data of the target vehicle is obtained, and the original laser radar point cloud data is converted into an obstacle binary matrix; the corresponding target collision detection strategy is determined according to the parking scene of the target vehicle to obtain the vehicle parameters corresponding to the target collision detection strategy, and the obstacle expansion radius of the target vehicle is calculated based on the vehicle parameters and the target collision detection strategy to calculate the corresponding structural elements through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy; based on the obstacle binary matrix and the structural data, an image expansion operation is performed to generate an obstacle map. Figure 2 Value image matrix, and use the obstacle map Figure 2 The corresponding obstacle map is constructed based on the value image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular border sampling strategy or the multi-circle strategy according to the current position, and the target vehicle is subjected to collision detection based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy. The present application generates an obstacle map and performs efficient and rapid collision detection by accurately describing the obstacles and selecting the appropriate collision detection form according to the scene. Thus, the problems that the prior art loses part of the passable area in narrow and complex scenes, and the use of the occupied grid to describe the map have high requirements on computing power and equipment are solved.

[0034] Specifically, Figure 1 A flowchart of a collision detection method based on an automatic parking obstacle map provided in an embodiment of the present application.

[0035] like Figure 1 As shown, the collision detection method based on the automatic parking obstacle map includes the following steps:

[0036] In step S101, the original laser radar point cloud data of the target vehicle is obtained, and the original laser radar point cloud data is converted into an obstacle binary matrix.

[0037] The embodiments of the present application can first acquire and process the original laser radar point cloud data of the target vehicle, and set the pixel map step size, so as to process the original laser radar point cloud data into matrix coordinates to obtain the corresponding obstacle binary matrix.

[0038] Optionally, in one embodiment of the present application, the laser radar raw point cloud data of the target vehicle is obtained, and the laser radar raw point cloud data is converted into an obstacle binary matrix, including: eliminating the ground point cloud in the three-dimensional laser radar of the laser radar raw point cloud data and the point cloud that does not meet the preset height and distance requirements to obtain the three-dimensional residual point cloud, and removing the height value of the three-dimensional residual point cloud to generate the corresponding first two-dimensional point cloud; eliminating the point cloud in the two-dimensional laser radar of the laser radar raw point cloud data that does not meet the preset distance requirement to obtain a second two-dimensional point cloud, and translating the first two-dimensional point cloud and the second two-dimensional point cloud to obtain a translated point set; constructing an obstacle binary matrix based on the translated point set and the preset pixel map step size.

[0039] It should be noted that for the three-dimensional laser radar in the original laser radar point cloud data, the embodiment of the present application can eliminate the ground point cloud, the point cloud whose height exceeds a certain range of the vehicle, and the point cloud far away from the vehicle to obtain the three-dimensional residual point cloud, and remove the height value of the three-dimensional residual point cloud to obtain the corresponding two-dimensional point cloud (i.e., the first two-dimensional point cloud); for the two-dimensional laser radar in the original laser radar point cloud data, the point cloud far away from the vehicle is eliminated to obtain the two-dimensional point cloud corresponding to the two-dimensional laser radar (i.e., the two-dimensional laser radar); further, the embodiment of the present application can translate the processed two-dimensional point cloud (i.e., the first two-dimensional point cloud and the second two-dimensional point cloud) as a whole, so that the minimum X and Y coordinates of the point set are 0, and the translation point set P = {[x 1 ,y 1 ];…;[x m ,y m ]}, containing n sets of two-dimensional points.

[0040] Afterwards, the embodiment of the present application sets the map step length L, and the maximum X-axis value of the translation point set P is x max , the maximum Y-axis value is y max , to establish an M*N all-0 matrix, where Then the coordinates of the i-th group of two-dimensional points in P on the matrix are: Let the matrix (X i , Y i ) position is 1, then the obstacle binary matrix Q can be obtained, where X i , Y iIt is a non-negative integer. [*] means rounding to the nearest integer.

[0041] Therefore, the embodiments of the present application process the original point cloud data of the laser radar to obtain an obstacle binary matrix, thereby providing reliable data support for the realization of automatic parking obstacle map generation and collision detection.

[0042] In step S102, a corresponding target collision detection strategy is determined according to the parking scenario of the target vehicle to obtain vehicle parameters corresponding to the target collision detection strategy, and an obstacle expansion radius of the target vehicle is calculated based on the vehicle parameters and the target collision detection strategy to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy.

[0043] Furthermore, the embodiment of the present application also needs to select a suitable collision detection strategy (i.e., target collision detection strategy) according to the parking scenario. The collision detection strategy is divided into a rectangular border sampling strategy and a multi-circle strategy. Secondly, the embodiment of the present application can obtain vehicle parameters. For the rectangular border sampling strategy, the sampling step size can be set accordingly to calculate the initial vehicle border sampling point coordinates (i.e., the vehicle border point set). For the multi-circle strategy, the embodiment of the present application can use W (the embodiment of the present application takes 10 circles as an example) circles of the same radius to completely surround the vehicle according to the vehicle parameters, such as Figure 2 As shown, the radius R and the center coordinates of the circle are calculated;

[0044] Afterwards, the embodiments of the present application can calculate the corresponding obstacle expansion radius based on the circle radius R, the center coordinates, the vehicle frame point set, and combined with the collision detection strategy, and calculate the corresponding structural element according to the obstacle expansion radius.

[0045] Optionally, in one embodiment of the present application, a corresponding target collision detection strategy is determined according to the parking scenario of the target vehicle to obtain vehicle parameters corresponding to the target collision detection strategy, and the obstacle expansion radius of the target vehicle is calculated based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy, including: when the target collision detection strategy is a rectangular border sampling strategy, determining a sampling step corresponding to the rectangular border sampling strategy, so as to perform a sampling operation according to the sampling step and the vehicle parameters to obtain a corresponding vehicle border point set; when the target collision detection strategy is a multi-circle strategy, dividing the vehicle rectangle corresponding to the target vehicle into multiple sub-rectangles according to the vehicle parameters and preset division requirements, and calculating the circumscribed circle radius and the center coordinates of each sub-rectangle in the multiple sub-rectangles, so as to construct a center two-dimensional point set corresponding to the multi-circle strategy based on the circumscribed circle radius and the center coordinates; based on the preset safety distance, the vehicle border point set and the center two-dimensional point set, calculating the obstacle expansion radius corresponding to the rectangular border sampling strategy or the multi-circle strategy, and calculating the corresponding structural element according to the obstacle expansion radius.

[0046] It should be noted that the embodiments of the present application can be divided into two forms of collision detection (i.e., two collision detection strategies). For narrow and complex scenes, the embodiments of the present application can use the precise collision detection strategy to make full use of the passable area, corresponding to the rectangular frame sampling strategy; for open scenes, the multi-circle strategy is used to improve the efficiency of collision detection.

[0047] In the actual implementation process, Figure 3 As shown, for the rectangular border sampling strategy, the embodiment of the present application can obtain vehicle parameters, including length, width, front axle, and rear axle positions, and represent the vehicle with a directed rectangle. For the rectangular border sampling strategy, the embodiment of the present application can set a sampling step d (the sampling step can be set to 0.1 meters), set the center of the rear axle of the vehicle to the origin, set the heading angle to 0, and sample a point in the rectangular border every step d, such as Figure 4 As shown, it is saved as a two-dimensional point set (i.e., the vehicle bounding box point set) A.

[0048] For the multi-circle strategy, the embodiment of the present application can place the center of the vehicle's rear axle at the origin, set the heading angle to 0, and evenly divide the vehicle rectangle into W small rectangles of the same size and consistent direction, calculate the circumscribed circle radius R and center coordinates of each small rectangle, and save the two-dimensional point set B of the center.

[0049] Afterwards, the embodiment of the present application can set a safety distance S. For the rectangular frame sampling strategy, the obstacle expansion radius is S, and the structural element is K. 1 *K 1 The all-1 matrix M 1 ,in For the multi-circle strategy, the obstacle expansion radius is S+R and the structural element is K. 2 *K 2 The all-1 matrix M 2 ,in [*] indicates rounding to the nearest integer.

[0050] In step S103, an image expansion operation is performed based on the obstacle binary matrix and the structure data to generate an obstacle map. Figure 2 Value image matrix, and use the obstacle map Figure 2 The corresponding obstacle map is constructed based on the value image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular border sampling strategy or the multi-circle strategy according to the current position, and collision detection is performed on the target vehicle based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy.

[0051] Furthermore, the embodiment of the present application further needs to regard the obstacle binary matrix as a binary image, and use the structure element to perform an image expansion operation to generate an obstacle map for collision detection. Figure 2 value image matrix; then, the embodiment of the present application can calculate the matrix coordinates of the points in the vehicle bounding box point set on the obstacle map under the rectangular bounding box sampling strategy according to the current position of the vehicle, or calculate the matrix coordinates of the centers of each enclosing circle on the obstacle map under the multi-circle strategy to obtain the corresponding matrix coordinate point set; finally, the embodiment of the present application can perform collision detection on the vehicle based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy.

[0052] Optionally, in one embodiment of the present application, an image dilation operation is performed based on the obstacle binary matrix and the structure data to generate an obstacle map. Figure 2 Value image matrix, and use the obstacle map Figure 2 The corresponding obstacle map is constructed by using the target collision detection strategy as the image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, including: using the structural element of the target collision detection strategy to perform an image expansion operation on the obstacle binary matrix to obtain the obstacle map corresponding to the target collision detection strategy. Figure 2 value image matrix; based on the current position of the target vehicle and the preset pixel map step size, the matrix coordinate point set of the vehicle border point set or the center two-dimensional point set on the obstacle binary image matrix is ​​calculated.

[0053] It should be noted that the embodiment of the present application may regard the obstacle binary matrix Q as a binary image. For the rectangular frame sampling method, the embodiment of the present application may use the structure element M 1 Perform image dilation on Q to obtain the obstacle map Figure 2Value image matrix E 1 ; For the multi-circle method, use the structural element M 2 Perform image dilation on Q to obtain the obstacle map Figure 2 Value image matrix E 2 .

[0054] Afterwards, for the rectangular frame sampling method, the embodiment of the present application can calculate the vehicle frame point set A in the obstacle binary image matrix E according to the current position of the vehicle and the pixel map step length L 1 The matrix coordinate point set T on A For the multi-circle method, the embodiment of the present application can calculate the center point set B in the obstacle binary image matrix E according to the current position of the vehicle and the pixel map step length L 2 The matrix coordinate point set T on B .

[0055] As a possible way to achieve this, Figure 5 is a schematic diagram of obstacle expansion, such as Figure 5 As shown in the figure, its structural element is a 21*21 all-1 matrix, which can expand the original obstacle point outward by 10 pixels.

[0056] Optionally, in one embodiment of the present application, collision detection is performed on the target vehicle based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy, including: when the target collision detection strategy is a rectangular border sampling strategy, determining whether all points in the matrix coordinate point set corresponding to the rectangular border sampling strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle border point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle border point set, then the target vehicle has not collided; when the target collision detection strategy is a multi-circle strategy, determining whether all points in the matrix coordinate point set corresponding to the multi-circle strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the center two-dimensional point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the center two-dimensional point set, then the target vehicle has not collided.

[0057] In the specific implementation process, the embodiment of the present application can select the corresponding obstacle map matrix according to the collision detection strategy. For the rectangular frame sampling strategy, if the matrix coordinate point set T A All points (i.e. all vehicle border points) are located in the obstacle binary image matrix E 1 If is not 1, no collision occurs. Figure 6 This is a schematic diagram of collision detection using a rectangular border sampling strategy. Figure 6 As shown in , there is a vehicle bounding box sampling point located at a position with an obstacle on the obstacle map matrix, so a collision occurs; for the multi-circle strategy, if the matrix coordinate point set T BAll points (i.e. the centers of all the circles enclosed by the points) are located in the obstacle binary image matrix E 2 If the position is not 1, no collision occurs, thus completing the collision detection of the vehicle position during automatic parking path planning.

[0058] According to the collision detection method based on automatic parking obstacle map proposed in the embodiment of the present application, the laser radar raw point cloud data of the target vehicle is obtained and converted into an obstacle binary matrix; the corresponding target collision detection strategy is determined according to the parking scene of the target vehicle to obtain the vehicle parameters corresponding to the target collision detection strategy, and the obstacle expansion radius of the target vehicle is calculated based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular frame sampling strategy and a multi-circle strategy; based on the obstacle binary matrix and the structural data, an image expansion operation is performed to generate an obstacle map. Figure 2 value image matrix, and use the obstacle map Figure 2 The corresponding obstacle map is constructed by using the value image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, and the target vehicle is subjected to collision detection based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy. The present application generates an obstacle map and performs efficient and rapid collision detection by accurately describing the obstacles and selecting the appropriate collision detection form according to the scene.

[0059] Secondly, a collision detection device based on an automatic parking obstacle map proposed in an embodiment of the present application is described with reference to the accompanying drawings.

[0060] Figure 7 It is a block diagram of a collision detection device based on an automatic parking obstacle map according to an embodiment of the present application.

[0061] like Figure 7 As shown, the collision detection device 10 based on the automatic parking obstacle map includes: a conversion module 100 , a calculation module 200 and a collision detection module 300 .

[0062] The conversion module 100 is used to obtain the original laser radar point cloud data of the target vehicle and convert the original laser radar point cloud data into an obstacle binary matrix.

[0063] The calculation module 200 is used to determine the corresponding target collision detection strategy according to the parking scenario of the target vehicle, to obtain the vehicle parameters corresponding to the target collision detection strategy, and to calculate the obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular border sampling strategy and a multi-circle strategy.

[0064] The collision detection module 300 is used to perform an image expansion operation based on the obstacle binary matrix and the structure data to generate an obstacle map. Figure 2 Value image matrix, and use the obstacle map Figure 2 The corresponding obstacle map is constructed based on the value image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular border sampling strategy or the multi-circle strategy according to the current position, and collision detection is performed on the target vehicle based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy.

[0065] Optionally, in one embodiment of the present application, the conversion module 100 includes: a first rejection unit, a second rejection unit, a translation unit and a construction unit.

[0066] Among them, the first elimination unit is used to eliminate the ground point cloud in the three-dimensional laser radar of the laser radar original point cloud data and the point cloud that does not meet the preset height and distance requirements to obtain the three-dimensional residual point cloud, and remove the height value of the three-dimensional residual point cloud to generate the corresponding first two-dimensional point cloud.

[0067] The second elimination unit is used to eliminate the point cloud that does not meet the preset distance requirement in the two-dimensional laser radar of the laser radar original point cloud data to obtain a second two-dimensional point cloud.

[0068] The translation unit is used to translate the first two-dimensional point cloud and the second two-dimensional point cloud to obtain a translation point set.

[0069] The construction unit is used to construct an obstacle binary matrix based on the translation point set and the preset pixel map step size.

[0070] Optionally, in one embodiment of the present application, the calculation module 200 includes: a sampling unit, a division unit and a calculation unit.

[0071] Among them, the sampling unit is used to determine the sampling step corresponding to the rectangular border sampling strategy when the target collision detection strategy is the rectangular border sampling strategy, so as to perform sampling operations according to the sampling step and vehicle parameters to obtain the corresponding vehicle border point set.

[0072] The division unit is used to divide the vehicle rectangle corresponding to the target vehicle into multiple sub-rectangles according to vehicle parameters and preset division requirements when the target collision detection strategy is a multi-circle strategy, and calculate the circumscribed circle radius and center coordinates of each sub-rectangle in the multiple sub-rectangles, so as to construct a two-dimensional point set of the center corresponding to the multi-circle strategy based on the circumscribed circle radius and the center coordinates.

[0073] The computing unit is used to calculate the obstacle expansion radius corresponding to the rectangular border sampling strategy or the multi-circle strategy based on the preset safety distance, the vehicle border point set and the circle center two-dimensional point set, and calculate the corresponding structural element according to the obstacle expansion radius.

[0074] Optionally, in one embodiment of the present application, the collision detection module 300 includes: an image expansion unit and a computing unit.

[0075] The image expansion unit is used to perform an image expansion operation on the obstacle binary matrix using the structural element of the target collision detection strategy to obtain the obstacle map corresponding to the target collision detection strategy. Figure 2 Value image matrix.

[0076] The computing unit is used to calculate the matrix coordinate point set of the vehicle border point set or the circle center two-dimensional point set on the obstacle binary image matrix based on the current position of the target vehicle and the preset pixel map step size.

[0077] Optionally, in one embodiment of the present application, the collision detection module 300 further includes: a first judgment unit and a second judgment unit.

[0078] Among them, the first judgment unit is used to judge whether all points in the matrix coordinate point set corresponding to the rectangular border sampling strategy are located in the non-1 position of the obstacle binary image matrix corresponding to the vehicle border point set when the target collision detection strategy is the rectangular border sampling strategy, wherein if all points are located in the non-1 position of the obstacle binary image matrix corresponding to the vehicle border point set, then the target vehicle has not collided.

[0079] The second judgment unit is used to judge whether all points in the matrix coordinate point set corresponding to the multi-circle strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the center two-dimensional point set when the target collision detection strategy is the multi-circle strategy, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the center two-dimensional point set, then the target vehicle has not collided.

[0080] It should be noted that the aforementioned explanation of the embodiment of the collision detection method based on the automatic parking obstacle map is also applicable to the collision detection device based on the automatic parking obstacle map of this embodiment, and will not be repeated here.

[0081] The collision detection device based on the automatic parking obstacle map proposed in the embodiment of the present application includes a conversion module 100, which is used to obtain the laser radar original point cloud data of the target vehicle and convert the laser radar original point cloud data into an obstacle binary matrix; a calculation module 200, which is used to determine the corresponding target collision detection strategy according to the parking scene of the target vehicle, so as to obtain the vehicle parameters corresponding to the target collision detection strategy, and calculate the obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy, so as to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular frame sampling strategy and a multi-circle strategy; a collision detection module 300, which is used to perform an image expansion operation based on the obstacle binary matrix and the structural data to generate an obstacle map. Figure 2 value image matrix, and use the obstacle map Figure 2 The corresponding obstacle map is constructed by using the value image matrix, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, and the target vehicle is subjected to collision detection based on the collision detection strategy corresponding to the matrix coordinate point set and the target collision detection strategy. The present application generates an obstacle map and performs efficient and rapid collision detection by accurately describing the obstacles and selecting the appropriate collision detection form according to the scene.

[0082] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0083] A memory 801 , a processor 802 , and a computer program stored in the memory 801 and executable on the processor 802 .

[0084] When the processor 802 executes the program, the collision detection method based on the automatic parking obstacle map provided in the above embodiment is implemented.

[0085] Furthermore, the electronic device further comprises:

[0086] The communication interface 803 is used for communication between the memory 801 and the processor 802 .

[0087] The memory 801 is used to store computer programs that can be executed on the processor 802 .

[0088] The memory 801 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0089] If the memory 801, the processor 802 and the communication interface 803 are implemented independently, the communication interface 803, the memory 801 and the processor 802 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0090] Optionally, in a specific implementation, if the memory 801, the processor 802 and the communication interface 803 are integrated on a chip, the memory 801, the processor 802 and the communication interface 803 can communicate with each other through an internal interface.

[0091] The processor 802 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0092] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned collision detection method based on the automatic parking obstacle map.

[0093] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned collision detection method based on the automatic parking obstacle map.

[0094] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0095] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0096] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0098] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0099] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0100] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0101] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A collision detection method based on automatic parking obstacle map, characterized in that: The following steps are involved: Acquire the original laser radar point cloud data of the target vehicle, and convert the original laser radar point cloud data into an obstacle binary matrix; Determining a corresponding target collision detection strategy according to the parking scene of the target vehicle to obtain vehicle parameters corresponding to the target collision detection strategy, and calculating an obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy to calculate a corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular frame sampling strategy and a multi-circle strategy; Based on the obstacle binary matrix and the structural data, an image dilation operation is performed to generate an obstacle map binary image matrix, and the obstacle map binary image matrix is ​​used to construct a corresponding obstacle map, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, and based on the matrix coordinate point set and the collision detection strategy corresponding to the target collision detection strategy, collision detection is performed on the target vehicle.

2. The method according to claim 1, characterized in that The step of acquiring the original laser radar point cloud data of the target vehicle and converting the original laser radar point cloud data into an obstacle binary matrix includes: Eliminating the ground point cloud in the three-dimensional laser radar of the laser radar original point cloud data and the point cloud that does not meet the preset height and distance requirements to obtain a three-dimensional residual point cloud, and removing the height value of the three-dimensional residual point cloud to generate a corresponding first two-dimensional point cloud; Eliminating point clouds that do not meet a preset distance requirement in the two-dimensional laser radar of the laser radar original point cloud data to obtain a second two-dimensional point cloud; Translating the first two-dimensional point cloud and the second two-dimensional point cloud to obtain a translation point set; The obstacle binary matrix is ​​constructed based on the translation point set and a preset pixel map step size.

3. The method according to claim 2, characterized in that The target collision detection strategy corresponding to the target vehicle is determined according to the parking scene of the target vehicle to obtain vehicle parameters corresponding to the target collision detection strategy, and the obstacle expansion radius of the target vehicle is calculated based on the vehicle parameters and the target collision detection strategy to calculate the corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular frame sampling strategy and a multi-circle strategy, including: When the target collision detection strategy is the rectangular frame sampling strategy, determining a sampling step corresponding to the rectangular frame sampling strategy, performing a sampling operation according to the sampling step and the vehicle parameters, so as to obtain a corresponding vehicle frame point set; When the target collision detection strategy is the multi-circle strategy, the vehicle rectangle corresponding to the target vehicle is divided into a plurality of sub-rectangles according to the vehicle parameters and the preset division requirements, and the circumscribed circle radius and the center coordinates of each of the plurality of sub-rectangles are calculated, so as to construct a two-dimensional point set of the center corresponding to the multi-circle strategy based on the circumscribed circle radius and the center coordinates; Based on the preset safety distance, the vehicle frame point set and the circle center two-dimensional point set, the obstacle expansion radius corresponding to the rectangular frame sampling strategy or the multi-circle strategy is calculated, and the corresponding structural element is calculated according to the obstacle expansion radius.

4. The method according to claim 3, characterized in that The image dilation operation is performed based on the obstacle binary matrix and the structural data to generate an obstacle map binary image matrix, and the obstacle map binary image matrix is ​​used to construct a corresponding obstacle map, and the current position of the target vehicle is obtained to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, including: Performing an image expansion operation on the obstacle binary matrix using the structural element of the target collision detection strategy to obtain a binary image matrix of the obstacle map corresponding to the target collision detection strategy; Based on the current position of the target vehicle and a preset pixel map step size, the matrix coordinate point set of the vehicle frame point set or the circle center two-dimensional point set on the obstacle binary image matrix is ​​calculated.

5. The method according to claim 4, characterized in that The collision detection strategy based on the matrix coordinate point set and the target collision detection strategy corresponding to the target collision detection strategy, performing collision detection on the target vehicle, comprises: When the target collision detection strategy is the rectangular frame sampling strategy, it is determined whether all points in the matrix coordinate point set corresponding to the rectangular frame sampling strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle frame point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the vehicle frame point set, then the target vehicle has not collided; When the target collision detection strategy is the multi-circle strategy, it is determined whether all points in the matrix coordinate point set corresponding to the multi-circle strategy are located at non-1 positions of the obstacle binary image matrix corresponding to the circle center two-dimensional point set, wherein if all points are located at non-1 positions of the obstacle binary image matrix corresponding to the circle center two-dimensional point set, then the target vehicle has not collided.

6. A collision detection device based on an automatic parking obstacle map, characterized in that: include: A conversion module, used to obtain the original laser radar point cloud data of the target vehicle and convert the original laser radar point cloud data into an obstacle binary matrix; a calculation module, configured to determine a corresponding target collision detection strategy according to a parking scenario of the target vehicle, to obtain vehicle parameters corresponding to the target collision detection strategy, and to calculate an obstacle expansion radius of the target vehicle based on the vehicle parameters and the target collision detection strategy, so as to calculate a corresponding structural element through the obstacle expansion radius, wherein the target collision detection strategy includes a rectangular frame sampling strategy and a multi-circle strategy; A collision detection module is used to perform an image expansion operation based on the obstacle binary matrix and the structural data to generate an obstacle map binary image matrix, and use the obstacle map binary image matrix to construct a corresponding obstacle map, and obtain the current position of the target vehicle to calculate the matrix coordinate point set of the target vehicle on the corresponding obstacle map under the rectangular frame sampling strategy or the multi-circle strategy according to the current position, and perform collision detection on the target vehicle based on the matrix coordinate point set and the collision detection strategy corresponding to the target collision detection strategy.

7. The device according to claim 6, characterized in that The conversion module comprises: A first elimination unit is used to eliminate the ground point cloud in the three-dimensional laser radar of the laser radar original point cloud data and the point cloud that does not meet the preset height and distance requirements to obtain a three-dimensional residual point cloud, and remove the height value of the three-dimensional residual point cloud to generate a corresponding first two-dimensional point cloud; A second elimination unit is used to eliminate the point cloud that does not meet the preset distance requirement in the two-dimensional laser radar of the laser radar original point cloud data to obtain a second two-dimensional point cloud; A translation unit, used for translating the first two-dimensional point cloud and the second two-dimensional point cloud to obtain a translation point set; A construction unit is used to construct the obstacle binary matrix based on the translation point set and a preset pixel map step size.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the collision detection method based on the automatic parking obstacle map as described in any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the collision detection method based on the automatic parking obstacle map as described in any one of claims 1-5.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the collision detection method based on the automatic parking obstacle map as described in any one of claims 1-5.

Images