Obstacle recognition method, device, equipment and storage medium based on unmanned vehicle

Abstract

The present application discloses an obstacle recognition method, device, equipment and storage medium based on an unmanned vehicle, wherein the method comprises: three-dimensionally rasterizing the body area of ​​the unmanned vehicle to obtain an initial body grid; obtaining the first point cloud data of the unmanned vehicle; matching the first point cloud data with the initial body grid to determine the target point cloud data falling into the initial body grid; optimizing the initial body grid based on the target point cloud data to generate a target body grid; obtaining the second point cloud data of the unmanned vehicle; matching the second point cloud data with the target body grid to determine the body point cloud data falling into the target body grid, as well as the vehicle point cloud data outside the target body grid; using the vehicle point cloud data outside the vehicle as obstacle point cloud data. The separation accuracy, efficiency and stability of the vehicle point cloud data and the vehicle point cloud data outside the vehicle can be improved, thereby improving the accuracy of obstacle recognition.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to an obstacle recognition method based on an unmanned vehicle, an obstacle recognition device based on an unmanned vehicle, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the development of information technology, unmanned driving technology is becoming more mature under the rapid development of artificial intelligence and the automotive industry. As a subdivision of unmanned driving technology, low-speed unmanned driving is mainly used in parks or semi-enclosed parks. Due to its low speed, low cost, high safety and mature technology, it is relatively easier to implement. At present, the products for low-speed unmanned driving scenarios mainly include unmanned sweepers, unmanned logistics distribution vehicles, unmanned patrol vehicles, etc.

[0003] Unmanned vehicles can sense the environment and their own status through sensors and avoid obstacles, so as to achieve autonomous driving. LiDAR is a sensor widely used in unmanned vehicles. LiDAR can obtain 3D point cloud data of the surrounding environment. In related technologies, unmanned vehicles need to perform point cloud pre-processing to detect obstacles, including segmentation steps such as removing the ground, removing the body, separating the below-ground, and separating the above-ground.

[0004] When segmenting the point cloud of the car body, the following two methods are used in related technologies to remove the point cloud of the car body:

[0005] Method 1 uses the body to construct a three-dimensional geometric polygonal area to remove part of the body point cloud. However, when the body geometry is irregular, filtering the point cloud has problems such as over-segmentation, under-segmentation, and mis-segmentation.

[0006] Method 2 uses the segmentation method of 3D grid occupancy, but the point cloud falls inside the vehicle body when the laser radar reflects light, trails, or absorbs points. Since the above phenomenon does not occur when filtering the point cloud of the vehicle body, it leads to the problem of false detection of obstacles. Summary of the invention

[0007] The present application provides an obstacle recognition method, device, equipment and storage medium based on an unmanned vehicle to solve the problems of inaccurate point cloud segmentation, false obstacle detection, etc. that occur when segmenting the point cloud of the vehicle body in the prior art.

[0008] In a first aspect, an embodiment of the present application provides an obstacle recognition method based on an unmanned vehicle, the method comprising:

[0009] Perform three-dimensional rasterization on the body area of ​​the unmanned vehicle to obtain an initial body grid;

[0010] Acquire first point cloud data of the unmanned vehicle;

[0011] Matching the first point cloud data with the initial vehicle body grid to determine target point cloud data falling into the initial vehicle body grid;

[0012] Optimizing the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid;

[0013] Acquire second point cloud data of the unmanned vehicle;

[0014] Matching the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid;

[0015] The external point cloud data is used as obstacle point cloud data.

[0016] In a second aspect, the embodiment of the present application further provides an obstacle recognition device based on an unmanned vehicle, the device comprising:

[0017] An initial vehicle body grid generation module is used to perform three-dimensional gridding on the vehicle body area of ​​the unmanned vehicle to obtain an initial vehicle body grid;

[0018] A first point cloud data acquisition module, used to acquire first point cloud data of the unmanned vehicle;

[0019] a target point cloud data determination module, used for matching the first point cloud data with the initial vehicle body grid to determine target point cloud data falling into the initial vehicle body grid;

[0020] An optimization module, used for optimizing the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid;

[0021] A second point cloud data acquisition module, used to acquire second point cloud data of the unmanned vehicle;

[0022] A point cloud segmentation module, used for matching the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid;

[0023] The obstacle point cloud determination module is used to use the external point cloud data as obstacle point cloud data.

[0024] In a third aspect, an embodiment of the present application further provides an electronic device, the electronic device comprising:

[0025] one or more processors;

[0026] a storage device for storing one or more programs,

[0027] When the one or more programs are executed by the one or more processors, the one or more processors implement the method of the first aspect described above.

[0028] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the method of the first aspect when executed by a processor.

[0029] The technical solution provided by this application has the following beneficial effects:

[0030] In this embodiment, the geometric body area of ​​the unmanned vehicle is three-dimensionally rasterized to obtain an initial body grid, which is then optimized and adjusted in combination with the point cloud data of the unmanned vehicle so that the newly obtained target body grid can more accurately represent the surface of the unmanned vehicle. The target body grid is then used to separate the point cloud data of the unmanned vehicle, thereby obtaining a more accurate separation result, improving the accuracy, efficiency, and stability of the separation of the body point cloud data and the external point cloud data, and thereby improving the accuracy of obstacle recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a flow chart of an obstacle recognition method embodiment based on an unmanned vehicle provided in Embodiment 1 of the present application;

[0032] Figure 2 is a schematic diagram of matching the first point cloud data with the initial vehicle body grid provided in the first embodiment of the present application;

[0033] Figure 3 This is a schematic diagram of a target point cloud data set provided in Example 1 of the present application;

[0034] Figure 4 It is a schematic diagram of obtaining a target vehicle body grid after filling an initial vehicle body grid provided in the first embodiment of the present application;

[0035] Figure 5 It is a structural block diagram of an obstacle recognition device based on an unmanned vehicle provided in Embodiment 2 of the present application;

[0036] Figure 6 It is a structural schematic diagram of an electronic device provided in Example 3 of the present application. DETAILED DESCRIPTION

[0037] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only the parts related to the present application, rather than all structures, are shown in the accompanying drawings.

[0038] Embodiment 1

[0039] Figure 1 This is a flowchart of an obstacle recognition method embodiment based on an unmanned vehicle provided in Embodiment 1 of the present application. This embodiment can be applied to the scene of body point cloud segmentation of a low-speed unmanned vehicle. This embodiment can be executed by an unmanned vehicle, for example, by a controller of the unmanned vehicle. Alternatively, this embodiment can also be executed by a server, and after the server detects an obstacle, it can send the point cloud data of the obstacle to the unmanned vehicle to guide the unmanned vehicle to perform obstacle avoidance processing. This embodiment does not limit the execution subject.

[0040] like Figure 1 As shown, this embodiment may include the following steps:

[0041] Step 110, three-dimensionally grid the body area of ​​the unmanned vehicle to obtain an initial body grid.

[0042] The three-dimensional rasterization process is a process of constructing the body area of ​​the unmanned vehicle into a three-dimensional initial body grid. Exemplarily, the three-dimensional rasterization can further be a three-dimensional cube rasterization, and the initial body grid can be composed of a plurality of grid units, wherein each grid unit can be a cube grid, and the side length of each grid unit is equal. In one embodiment, in order to facilitate subsequent matching with point cloud data, the side length of each grid unit can be less than a preset laser ranging accuracy value.

[0043] In one embodiment, step 110 may further include the following steps:

[0044] Step 110 - 1 , determining the size of the body area of ​​the unmanned vehicle.

[0045] In one implementation, the unmanned vehicle has a default size of the vehicle body area, which can be recorded in the relevant configuration data pre-configured for the unmanned vehicle. Therefore, the size of its vehicle body area can be read from the configuration data of the unmanned vehicle. Exemplarily, the size of the vehicle body area may include: vehicle body length, vehicle body width, and vehicle body height.

[0046] Specifically, the configuration data of the unmanned vehicle will include the size of its body area. By reading the configuration data, the size of the body area can be found from the configuration data. For example, the size of the body area can be expressed as: L*W*H, where L is the body length, W is the body width, and H is the body height.

[0047] In other embodiments, it is also possible to obtain an image of the unmanned vehicle captured by an image sensor, then identify the vehicle body area from the image, and measure the length, width, and height of the vehicle body area to obtain the size of the vehicle body area.

[0048] Step 110 - 2 , constructing a three-dimensional coordinate system, and constructing a spatial region of the vehicle body region in the three-dimensional coordinate system according to the size of the vehicle body region.

[0049] In one implementation, the three-dimensional coordinate system can be constructed using the center point of the rear wheel axle in the vehicle body region as the origin, wherein the center point of the rear wheel axle in the vehicle body region can also be obtained from the configuration data.

[0050] In other embodiments, the position of the center point of the rear wheel axle in the vehicle body area may also be identified from the image containing the unmanned vehicle, which is not limited in this embodiment.

[0051] The three-dimensional coordinate system may be a right-handed coordinate system.

[0052] After the three-dimensional coordinate system is constructed, a spatial region of the vehicle body region can be constructed in the three-dimensional coordinate system, and the spatial region can be used to describe the geometric region of the unmanned vehicle. In one embodiment, the above-mentioned construction of the spatial region of the vehicle body region in the three-dimensional coordinate system according to the size of the vehicle body region can further include the following steps:

[0053] Step 110-2-1, based on the vehicle body length, determine the maximum length position and the minimum length position corresponding to the vehicle body length in the x-axis of the three-dimensional coordinate system.

[0054] During implementation, the minimum length position on the x-axis may be preset, and the minimum length position may be the position where the origin is located, or a position at a certain length away from the origin, which is not limited in this embodiment.

[0055] After the minimum length position is determined, the minimum length position can be used as the starting point, the body length can be used as the length to determine the body length line segment, and then the end point of the body length line segment can be used as the reference point, and then the position at a preset distance after the reference point can be used as the maximum length position. In this way, the length of the line segment determined by the minimum length position and the maximum length position will be slightly greater than the body length.

[0056] Step 110-2-2, based on the width of the vehicle body, determine the maximum width position and the minimum width position corresponding to the width of the vehicle body in the y-axis of the three-dimensional coordinate system.

[0057] During implementation, the minimum width position on the y-axis may be preset, and the minimum width position may be the position of the origin, or a position a certain length away from the origin, which is not limited in this embodiment.

[0058] After the minimum width position is determined, the minimum width position can be used as the starting point and the body width as the length to determine the body width line segment, and then the end point of the body width line segment is used as the reference point, and then the position at a preset distance after the reference point is used as the maximum width position. In this way, the length of the line segment determined by the minimum width position and the maximum width position will be slightly greater than the body width.

[0059] Step 110-2-3, based on the vehicle body height, determine the maximum height position and the minimum height position corresponding to the vehicle body height in the z-axis of the three-dimensional coordinate system.

[0060] During implementation, the minimum height position on the z-axis may be preset, and the minimum height position may be the position of the origin, or a position a certain length away from the origin, which is not limited in this embodiment.

[0061] After the minimum height position is determined, the minimum height position can be used as the starting point and the vehicle height as the length to determine the vehicle height line segment, and then the end point of the vehicle height line segment is used as the reference point, and then the position at a preset distance after the reference point is used as the maximum height position. In this way, the length of the line segment determined by the minimum height position and the maximum height position will be slightly greater than the vehicle height.

[0062] Step 110-2-4, constructing a spatial area of ​​the vehicle body area according to the maximum length position, the minimum length position, the maximum width position, the minimum width position, the maximum height position and the minimum height position.

[0063] Specifically, after the maximum position and the minimum position in each direction in the three-dimensional coordinate system are determined, the geometric area of ​​the vehicle body area, that is, the space area, can be constructed based on the maximum position and the minimum position in each direction.

[0064] Step 110 - 3 , performing rasterization processing on the spatial region to generate an initial vehicle body grid.

[0065] The initial vehicle body grid obtained after rasterizing the spatial region includes a plurality of cubic grid units, the side lengths of each grid unit are equal, and the side lengths of each grid unit are less than a preset laser ranging accuracy value.

[0066] Step 120, obtaining first point cloud data of the unmanned vehicle.

[0067] Specifically, point cloud data refers to a set of vectors in a three-dimensional coordinate system. These vectors are usually expressed in the form of x, y, and z three-dimensional coordinates to represent the outer surface shape of an object. In one embodiment, the point cloud data may be laser radar point cloud data, which is point cloud data obtained after the laser radar scans the unmanned vehicle. Therefore, the point cloud data of the unmanned vehicle can be obtained from the laser radar. In order to distinguish it from the subsequent point cloud data, the point cloud data is recorded as the first point cloud data here.

[0068] Step 130 : Matching the first point cloud data with the initial vehicle body grid to determine target point cloud data falling into the initial vehicle body grid.

[0069] In this step, the first point cloud data can be matched to the three-dimensional coordinate system where the initial vehicle body grid is located, and then it is determined whether the position of each first point cloud data is occupied by a grid unit. If the position of a certain first point cloud data is occupied by a grid unit, the first point cloud data is determined to be the target point cloud data falling into the initial vehicle body grid. Otherwise, if the position of a certain first point cloud data is not occupied by a grid unit, it is determined that the first point cloud data is not the target point cloud data. Falling into the initial vehicle body grid includes falling on the surface or inside of the initial vehicle body grid.

[0070] For example, Figure 2 The first point cloud data of the unmanned vehicle is matched with the initial vehicle body grid, and the first point cloud data falling within the initial vehicle body grid A frame is determined as the target point cloud data. The target point cloud data is shown in FIG. Figure 3 shown.

[0071] Step 140: Optimize the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid.

[0072] In order to more accurately separate the body point cloud (i.e., body point cloud) from the point cloud outside the body, and improve the accuracy of obstacle recognition. After determining the target point cloud data that falls into the initial body grid, this step can optimize and adjust the initial body grid through the target point cloud data to obtain the target body grid. Compared with the original initial body grid, the target body grid can more accurately express the surface of the unmanned vehicle.

[0073] In one embodiment, step 140 may further include the following steps:

[0074] Step 140 - 1 , project each of the target point cloud data onto the z-axis to obtain a corresponding projection line.

[0075] Step 140 - 2 , respectively translate each of the projection lines along the z-axis to the position with the minimum height of the z-axis, and perform grid filling on the plane generated by the translation of each of the projection lines to obtain a target vehicle body grid.

[0076] Specifically, assume that the point cloud set composed of the target point cloud data is recorded as P1. For any target point cloud data in P1, it is projected onto the z-axis, and the projection line is assumed to be a. Then a is translated along the z-axis to the minimum height position Zmin of the z-axis, and then all the planes covered by the translation line a during the translation process are grid-filled. Among them, grid filling refers to filling the above-generated plane with grid units of a set size.

[0077] After all target point cloud data in P1 are processed, the generated target vehicle body grid can be Figure 4 shown.

[0078] It should be noted that the above embodiment uses the z-axis as the reference for projection and translation. In actual processing, the x-axis or y-axis may also be used for projection and translation. The projection and translation and subsequent grid filling principles are similar to those of the z-axis. This embodiment does not limit this.

[0079] By grid-filling the plane generated by translation through projection and translation, the filling area can be expanded to obtain a more accurate vehicle body grid.

[0080] Step 150: Acquire second point cloud data of the unmanned vehicle.

[0081] Step 160 : Matching the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid.

[0082] After obtaining the target vehicle grid, the point cloud data of the unmanned vehicle can be obtained in real time, recorded as the second point cloud data. Then the second point cloud data is matched with the target vehicle grid in the same three-dimensional coordinate system, so as to determine the point cloud data falling into the target vehicle grid as the vehicle point cloud data. In addition, the point cloud data outside the target vehicle grid is determined as the vehicle point cloud data.

[0083] In one implementation, the above-mentioned matching process may include: determining whether the location of each second point cloud data is occupied by a grid unit; if so, determining that the second point cloud data is a vehicle body point cloud data falling into the target vehicle body grid; if not, determining that the second point cloud data is an external vehicle point cloud data outside the target vehicle body grid.

[0084] Step 170: Use the external point cloud data as obstacle point cloud data.

[0085] After separating the vehicle body point cloud data and the vehicle exterior point cloud data, the vehicle body point cloud data can be filtered out, and the vehicle exterior point cloud data can be used as obstacle point cloud data. The obstacle point cloud data is used to guide the unmanned vehicle to perform obstacle avoidance processing.

[0086] In this embodiment, the geometric body area of ​​the unmanned vehicle is three-dimensionally rasterized to obtain an initial body grid, and the initial body grid is filled and adjusted in combination with the point cloud data of the unmanned vehicle, so that the newly obtained target body grid can more accurately represent the surface of the unmanned vehicle. The target body grid is used to separate the point cloud data of the unmanned vehicle, and a more accurate separation result can be obtained, thereby improving the separation accuracy, efficiency, and stability of the body point cloud data and the external point cloud data, thereby improving the accuracy of obstacle recognition.

[0087] Embodiment 2

[0088] Figure 5 The structural block diagram of an obstacle recognition device based on an unmanned vehicle provided in the second embodiment of the present application may include the following modules:

[0089] An initial vehicle body grid generation module 510 is used to perform three-dimensional gridding on the vehicle body area of ​​the unmanned vehicle to obtain an initial vehicle body grid;

[0090] A first point cloud data acquisition module 520, used to acquire first point cloud data of the unmanned vehicle;

[0091] A target point cloud data determination module 530 is used to match the first point cloud data with the initial vehicle body grid to determine target point cloud data falling into the initial vehicle body grid;

[0092] An optimization module 540, configured to optimize the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid;

[0093] A second point cloud data acquisition module 550, used to acquire second point cloud data of the unmanned vehicle;

[0094] A point cloud segmentation module 560 is used to match the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid;

[0095] The obstacle point cloud determination module 570 is used to use the external point cloud data as obstacle point cloud data.

[0096] In one embodiment, the initial vehicle body grid generation module 510 may further include the following submodules:

[0097] A vehicle body size determination submodule, used to determine the size of the vehicle body area of ​​the unmanned vehicle;

[0098] A space region construction submodule, used to construct a three-dimensional coordinate system, and to construct a space region of the vehicle body region in the three-dimensional coordinate system according to the size of the vehicle body region;

[0099] The rasterization processing submodule is used to perform rasterization processing on the spatial area to generate an initial vehicle body grid, wherein the initial vehicle body grid includes a plurality of cubic grid units, each grid unit has an equal side length, and the side length is less than a preset laser ranging accuracy value.

[0100] In one embodiment, the vehicle body size determination submodule is specifically used for:

[0101] The size of the vehicle body area is read from configuration data, wherein the configuration data is pre-configured data for the unmanned vehicle.

[0102] In one embodiment, the spatial region construction submodule is specifically used for:

[0103] Obtaining the center point position of the rear wheel axle in the vehicle body area;

[0104] A three-dimensional coordinate system is constructed by taking the center point of the rear wheel axle as the origin.

[0105] In one embodiment, the dimensions of the vehicle body region include: vehicle body length, vehicle body width and vehicle body height;

[0106] The spatial region construction submodule is also used for:

[0107] According to the vehicle body length, determining a maximum length position and a minimum length position corresponding to the vehicle body length in the x-axis of the three-dimensional coordinate system;

[0108] According to the vehicle body width, determining a maximum width position and a minimum width position corresponding to the vehicle body width in the y-axis of the three-dimensional coordinate system;

[0109] According to the vehicle body height, determining a maximum height position and a minimum height position corresponding to the vehicle body height in the z-axis of the three-dimensional coordinate system;

[0110] A spatial region of the vehicle body region is constructed according to the maximum length position, the minimum length position, the maximum width position, the minimum width position, the maximum height position, and the minimum height position.

[0111] In one embodiment, the optimization module 540 is specifically used to:

[0112] Projecting each of the target point cloud data onto the z-axis to obtain a corresponding projection line;

[0113] Each of the projection lines is translated along the z-axis to the position with the minimum height of the z-axis, and a plane generated by the translation of each of the projection lines is grid-filled to obtain a target vehicle body grid.

[0114] In one embodiment, the point cloud segmentation module 560 is specifically used to:

[0115] Determine whether the position where each second point cloud data is located is occupied by a grid unit;

[0116] If yes, determining that the second point cloud data is the vehicle body point cloud data falling into the target vehicle body grid;

[0117] If not, it is determined that the second point cloud data is external point cloud data outside the target vehicle body grid.

[0118] An obstacle recognition device based on an unmanned vehicle provided in an embodiment of the present application can execute an obstacle recognition method based on an unmanned vehicle in Embodiment 1 of the present application, and has functional modules and beneficial effects corresponding to the execution method.

[0119] Embodiment 3

[0120] Figure 6 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present application is shown in FIG. Figure 6 As shown, the electronic device includes a processor 610, a memory 620, an input device 630, and an output device 640; the number of the processor 610 in the electronic device can be one or more. Figure 6 A processor 610 is taken as an example; the processor 610, the memory 620, the input device 630 and the output device 640 in the electronic device can be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0121] The memory 620 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as the program instructions / modules corresponding to the above-mentioned embodiment 1 in the embodiments of the present application. The processor 610 executes various functional applications and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 620, that is, implements the method mentioned in the above-mentioned method embodiment 1.

[0122] The memory 620 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 620 may further include a memory remotely arranged relative to the processor 610, and these remote memories may be connected to the device / terminal / server via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0123] The input device 630 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the electronic device. The output device 640 may include a display device such as a display screen.

[0124] Embodiment 4

[0125] Embodiment 4 of the present application also provides a storage medium containing computer executable instructions, and the computer executable instructions are used to execute the method of embodiment 1 of the above method when executed by a computer processor.

[0126] Of course, the storage medium containing computer executable instructions provided in the embodiments of the present application is not limited to the operations of the method described above, but can also execute related operations in the method provided in any embodiment of the present application.

[0127] Embodiment 5

[0128] Embodiment 5 of the present application further provides a computer program product, which includes computer executable instructions, and the computer executable instructions are used to execute the method of embodiment 1 of the above method when executed by a computer processor.

[0129] Of course, the computer program product provided by the embodiment of the present application, whose computer executable instructions are not limited to the method operations described above, can also execute related operations in the method provided by any embodiment of the present application.

[0130] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present application can be implemented with the help of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application can be essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for an electronic device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0131] It is worth noting that in the embodiment of the above-mentioned device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application.

[0132] Note that the above are only preferred embodiments of the present application and the technical principles used. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. An obstacle recognition method based on an unmanned vehicle, characterized in that: The method comprises: Perform three-dimensional rasterization on the body area of ​​the unmanned vehicle to obtain an initial body grid; Acquire first point cloud data of the unmanned vehicle; Matching the first point cloud data with the initial vehicle body grid to determine target point cloud data falling into the initial vehicle body grid; Optimizing the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid; Acquire second point cloud data of the unmanned vehicle; Matching the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid; The external point cloud data is used as obstacle point cloud data.

2. The method according to claim 1, characterized in that: The three-dimensional gridding of the vehicle body area of ​​the unmanned vehicle to obtain an initial vehicle body grid includes: Determine the size of the body area of ​​the unmanned vehicle; Constructing a three-dimensional coordinate system, and constructing a spatial region of the vehicle body region in the three-dimensional coordinate system according to the size of the vehicle body region; The spatial region is subjected to a rasterization process to generate an initial vehicle body grid, wherein the initial vehicle body grid includes a plurality of cubic grid units, the side lengths of each grid unit are equal, and the side lengths are less than a preset laser ranging accuracy value.

3. The method according to claim 2, characterized in that Determining the size of the body area of ​​the unmanned vehicle includes: The size of the vehicle body area is read from configuration data, wherein the configuration data is pre-configured data for the unmanned vehicle.

4. The method according to claim 2, characterized in that: The constructing of the three-dimensional coordinate system comprises: Obtaining the center point position of the rear wheel axle in the vehicle body area; A three-dimensional coordinate system is constructed by taking the center point of the rear wheel axle as the origin.

5. The method according to claim 2, 3 or 4, characterized in that: The dimensions of the vehicle body area include: vehicle body length, vehicle body width and vehicle body height; The step of constructing a spatial region of the vehicle body region in the three-dimensional coordinate system according to the size of the vehicle body region includes: According to the vehicle body length, determining a maximum length position and a minimum length position corresponding to the vehicle body length in the x-axis of the three-dimensional coordinate system; According to the vehicle body width, determining a maximum width position and a minimum width position corresponding to the vehicle body width in the y-axis of the three-dimensional coordinate system; According to the vehicle body height, determining a maximum height position and a minimum height position corresponding to the vehicle body height in the z-axis of the three-dimensional coordinate system; A spatial region of the vehicle body region is constructed according to the maximum length position, the minimum length position, the maximum width position, the minimum width position, the maximum height position, and the minimum height position.

6. The method according to claim 5, characterized in that The step of optimizing the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid includes: Projecting each of the target point cloud data onto the z-axis to obtain a corresponding projection line; Each of the projection lines is translated along the z-axis to the position with the minimum height of the z-axis, and a plane generated by the translation of each of the projection lines is grid-filled to obtain a target vehicle body grid.

7. The method according to any one of claims 1 to 4, characterized in that: The matching of the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid includes: Determine whether the position where each second point cloud data is located is occupied by a grid unit; If yes, determining that the second point cloud data is the vehicle body point cloud data falling into the target vehicle body grid; If not, it is determined that the second point cloud data is external point cloud data outside the target vehicle body grid.

8. An obstacle recognition device based on an unmanned vehicle, characterized in that: The device comprises: An initial vehicle body grid generation module is used to perform three-dimensional gridding on the vehicle body area of ​​the unmanned vehicle to obtain an initial vehicle body grid; A first point cloud data acquisition module, used to acquire first point cloud data of the unmanned vehicle; a target point cloud data determination module, used for matching the first point cloud data with the initial vehicle body grid to determine target point cloud data falling into the initial vehicle body grid; An optimization module, used for optimizing the initial vehicle body grid based on the target point cloud data to generate a target vehicle body grid; A second point cloud data acquisition module, used to acquire second point cloud data of the unmanned vehicle; A point cloud segmentation module, used for matching the second point cloud data with the target vehicle body grid to determine the vehicle body point cloud data falling into the target vehicle body grid and the vehicle external point cloud data outside the target vehicle body grid; The obstacle point cloud determination module is used to use the external point cloud data as obstacle point cloud data.

9. An electronic device, characterized in that: The electronic device comprises: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images