A radar point cloud data downsampling processing method and related equipment

Abstract

The present invention discloses a radar point cloud data downsampling processing method and related equipment. The method comprises: obtaining downsampled boundary area point cloud data of a target vehicle, wherein the downsampled boundary area point cloud data is obtained by performing boundary segmentation processing on the original point cloud data obtained from the target vehicle based on a preset downsampled boundary width, a downsampled boundary downsampling coefficient, and a perception internal downsampling coefficient; performing height segmentation on the downsampled boundary area point cloud data according to a first preset height threshold to obtain first area point cloud data and second area point cloud data, and performing target recognition on the first area point cloud data and the second area point cloud data to obtain target object information. The radar point cloud data downsampling processing method provided in the present application can effectively reduce the number of points in the point cloud to ensure the efficiency of subsequent processing, and maintain the point cloud features to ensure the accuracy of subsequent clustering and other operations by performing downsampling operations at different scales within the point cloud space.

Description

Technical Field

[0001] This specification relates to the field of signal processing, and more specifically, to a radar point cloud data downsampling processing method and related equipment. Background Art

[0002] The number of raw point clouds from LiDAR is huge. In order to improve processing efficiency, the raw point clouds are often downsampled (also known as downsampling) to reduce the amount of point cloud data, thereby improving the efficiency of subsequent point cloud processing. Common point cloud downsampling methods include uniform downsampling, random downsampling, voxel downsampling, etc. Although the implementation principles and effects of various methods are different, the processing of point cloud space is equal. For example, voxel downsampling divides the point cloud space into tightly arranged grids, and uses the center of gravity of the point cloud in each grid as the downsampling output. Compared with the method of using the geometric center of the grid as the downsampling output, although the sampling within the grid of voxel downsampling is based on the local spatial distribution of the point cloud, the processing of the point cloud space is uniform in the overall grid generation process, that is, the grid size and spatial distribution are uniform.

[0003] This uniform spatial processing scheme results in a single downsampling scale, which is difficult to determine. If the grid size is too small, many points will remain after downsampling, making the downsampling effect ineffective. If the grid size is too large, point cloud features will be easily filtered out, affecting subsequent spatial clustering and other processing results, increasing the risk of missed targets. This problem is particularly prominent when the spatial range is large and the clustered targets are small. Summary of the Invention

[0004] The Summary of the Invention introduces a series of simplified concepts that will be further described in the Detailed Description of the Invention. The Summary of the Invention is not intended to limit the key features and essential features of the claimed technical solution, nor is it intended to determine the scope of protection of the claimed technical solution.

[0005] In order to provide downsampled point cloud data that better meets recognition requirements, in a first aspect, the present invention proposes a radar point cloud data downsampling processing method, the method comprising:

[0006] Obtaining downsampled boundary area point cloud data of the target vehicle, wherein the downsampled boundary area point cloud data is obtained by performing boundary segmentation processing on the original point cloud data obtained from the target vehicle based on a preset downsampled boundary width, a downsampled boundary downsampling coefficient, and a perceived internal downsampling coefficient;

[0007] The downsampled boundary area point cloud data is highly segmented according to a first preset height threshold to obtain first area point cloud data and second area point cloud data, wherein the first area point cloud data is point cloud data of an area corresponding to the first preset height threshold whose height is lower than or equal to the first preset height threshold, a first downsampling coefficient is set corresponding to the first area, the second area point cloud data is point cloud data corresponding to a height higher than the first preset height threshold, a second downsampling coefficient is set corresponding to the second area point cloud data, the perception internal downsampling coefficient is greater than the downsampling boundary downsampling coefficient, the downsampling boundary downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient;

[0008] Target recognition is performed on the first area point cloud data and the second area point cloud data to obtain target object information.

[0009] Optionally, the target object information includes object type information of the target object;

[0010] The above method further includes:

[0011] A third downsampling coefficient of the corresponding area of ​​the target object is determined based on the object type information.

[0012] Optionally, the determining of the third downsampling coefficient of the corresponding area of ​​the target object based on the object type information includes:

[0013] When the height of the target object is less than or equal to the first preset height threshold, the third downsampling coefficient corresponding to the target object is set to the first downsampling coefficient;

[0014] and / or,

[0015] When the height of the target object is greater than the first preset height threshold and the length is greater than the length threshold, or the height of the target object is greater than the first preset height threshold or the width is greater than the width threshold, the third downsampling coefficient corresponding to the target object is set to the perceptual internal downsampling coefficient;

[0016] and / or,

[0017] When the height of the target object is greater than the first preset height threshold and the length is less than or equal to the length threshold, the third downsampling coefficient corresponding to the target object is set as the downsampling coefficient of the point cloud in the non-low area in the downsampling boundary;

[0018] and / or,

[0019] When the height of the target object is greater than the first preset height threshold and the width is less than the width threshold, the third downsampling coefficient corresponding to the target object is set as the downsampling coefficient of the point cloud in the non-low area in the downsampling boundary.

[0020] Optionally, the above method further includes:

[0021] In the case where there is an overlapping area between the corresponding areas of the plurality of target objects, a smaller downsampling coefficient is selected among the corresponding areas of the overlapping target objects to perform downsampling processing on the overlapping area.

[0022] Optionally, the above method further includes:

[0023] When a smaller target object overlaps behind a larger target object, obtaining a movement pattern of the smaller target object;

[0024] Downsampling processing is performed in the overlapping area based on the movement rule and the smaller downsampling coefficient.

[0025] Optionally, the above method further includes:

[0026] In the case where the target object is a moving object, obtaining a moving speed of the moving object;

[0027] A fourth downsampling coefficient of the area corresponding to the moving object is determined based on the moving speed.

[0028] Optionally, the above method further includes:

[0029] In the case where the target object is a moving object, obtaining a moving direction of the moving object;

[0030] In a case where the moving direction of the object does not overlap with the moving direction of the target vehicle, the perception internal downsampling coefficient is set as the downsampling coefficient of the area corresponding to the moving object.

[0031] In a second aspect, the present invention further provides a radar point cloud data downsampling processing device, comprising:

[0032] an acquisition unit, configured to acquire downsampled boundary area point cloud data of the target vehicle, wherein the downsampled boundary area point cloud data is acquired by performing boundary segmentation processing on the original point cloud data acquired from the target vehicle based on a preset downsampled boundary width, a downsampled boundary downsampling coefficient, and a perceived internal downsampling coefficient;

[0033] a segmentation unit, configured to perform height segmentation on the downsampled boundary area point cloud data according to a first downsampling coefficient and a second downsampling coefficient based on a first preset height threshold to obtain first area point cloud data and second area point cloud data, wherein the first area point cloud data is point cloud data of an area corresponding to a height lower than or equal to the first preset height threshold, the second area point cloud data is point cloud data corresponding to a height higher than the first preset height threshold, the perception internal downsampling coefficient is greater than the downsampling boundary downsampling coefficient, the downsampling boundary downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient;

[0034] The recognition unit is used to perform target recognition on the first area point cloud data and the second area point cloud data to obtain target object information.

[0035] In a third aspect, an electronic device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to implement the steps of the radar point cloud data downsampling processing method according to any one of the first aspects described above when executing the computer program stored in the memory.

[0036] In a fourth aspect, the present invention further proposes a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the radar point cloud data downsampling processing method according to any one of the above items in the first aspect is implemented.

[0037] In summary, the radar point cloud data downsampling processing method of the embodiment of the present application includes: obtaining downsampled boundary area point cloud data of the target vehicle, wherein the downsampled boundary area point cloud data is obtained by performing boundary segmentation processing on the original point cloud data obtained by the target vehicle based on a preset downsampled boundary width, a downsampled boundary downsampling coefficient, and a perceived internal downsampling coefficient; according to a first preset height threshold, the downsampled boundary area point cloud data is highly segmented according to a first downsampling coefficient and a second downsampling coefficient to obtain first area point cloud data and second area point cloud data, wherein the first area point cloud data is point cloud data of an area corresponding to a height lower than or equal to the first preset height threshold, the second area point cloud data is point cloud data corresponding to a height higher than the first preset height threshold, the perceived internal downsampling coefficient is greater than the downsampling boundary downsampling coefficient, the downsampling boundary downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient; target recognition is performed on the first area point cloud data and the second area point cloud data to obtain target object information. The radar point cloud data downsampling processing method provided in the embodiment of the present application performs boundary segmentation on the original point cloud data based on a preset downsampling boundary width and a downsampling boundary downsampling coefficient, and performs initial downsampling processing to obtain point cloud data in the downsampling boundary area. Then, based on a first preset height threshold, the point cloud data in the downsampling boundary area is secondary downsampled according to a first downsampling coefficient and a second downsampling. By using different scales for downsampling operations in the point cloud space, the number of points in the point cloud can be effectively reduced to ensure the efficiency of subsequent processing, while maintaining the point cloud features to ensure the accuracy of subsequent clustering and other operations.

[0038] The radar point cloud data downsampling processing method of the present invention, and other advantages, objectives, and features of the present invention will be partially reflected in the following description and partially understood by those skilled in the art through research and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present description. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0040] Figure 1 A schematic flow chart of a radar point cloud data downsampling processing method provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of a range for obtaining radar point cloud data for a vehicle provided in an embodiment of the present application;

[0042] Figure 3 A schematic diagram of the positional relationship between a vehicle and a target object provided in an embodiment of the present application;

[0043] Figure 4 A flowchart of another radar point cloud data downsampling processing method provided in an embodiment of the present application;

[0044] Figure 5 A schematic diagram of the structure of a radar point cloud data downsampling processing device provided in an embodiment of the present application;

[0045] Figure 6 A schematic diagram of the structure of an electronic device for downsampling and processing radar point cloud data provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] The radar point cloud data downsampling processing method provided in the embodiment of the present application performs boundary segmentation on the original point cloud data based on a preset downsampling boundary and boundary width into an outer boundary area, a downsampled boundary area, and an inner area. The downsampled boundary point cloud data is then further segmented into a low boundary area and a non-low boundary area based on a first preset height threshold. By using different scales for downsampling operations within the different segmented point cloud spaces, the number of points in the point cloud can be effectively reduced to ensure the efficiency of subsequent processing, while maintaining the point cloud features to ensure the accuracy of subsequent clustering and other operations.

[0047] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or devices. The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present application, not all of the embodiments.

[0048] It should be noted that for ease of explanation, the following assumes a positive correlation between the downsampling coefficient and the downsampling rate. That is, larger downsampling coefficients and higher downsampling rates result in more points being filtered out and fewer points being retained. Smaller downsampling coefficients and lower downsampling rates result in fewer points being filtered out and more points being retained. This assumption does not constitute a limitation of the technology. A downsampling boundary is set. Points outside the boundary are not downsampled due to their distance and sparseness. Points within the boundary are downsampled.

[0049] See also Figure 1 , which is a flowchart of a radar point cloud data downsampling processing method provided in an embodiment of the present application, which may specifically include:

[0050] S110, obtaining downsampled boundary area point cloud data of the target vehicle, wherein the downsampled boundary area point cloud data is obtained by performing boundary segmentation processing on original point cloud data obtained from the target vehicle based on a preset downsampled boundary and boundary width;

[0051] For example, the target vehicle can obtain the original point cloud data around the road through the radar sensor installed on the vehicle. The amount of original point cloud data is huge and needs to be downsampled before it can be used for obstacle recognition. The method proposed in this application first performs boundary segmentation processing on the original point cloud data according to the preset downsampling boundary and boundary width Δd, such as Figure 2 Figure 2 shows the region of point cloud data captured by multiple radar sensors on a vehicle. The raw point cloud data is segmented using a preset downsampling boundary and a boundary width Δd. The region within the downsampling boundary is called the boundary region. The point cloud within the boundary region is denoted as P1, and the point cloud within the boundary region is denoted as P2. The external point cloud outside the boundary region is sparse due to its distance and is not downsampled. The boundary downsampling coefficient is smaller than the internal downsampling coefficient, meaning that the data density of the downsampled boundary point cloud is greater than that of the internal perception point cloud. It should be noted that the downsampling boundary is generally a target area within the sensor's perception range, defined based on the vehicle type, road, and project requirements. To ensure timely detection of obstacles entering the perception region, the point cloud space near the perception region boundary is designated as the key perception region, and a smaller sampling grid is used. The boundary perception region distance threshold is set to Δd (Δd can be set to n·a0, where n is a positive integer and a0 is the initial downsampling coefficient). The region Δd inward from the downsampling boundary is considered the boundary region.

[0052] S120, performing height segmentation on the downsampled boundary area point cloud data according to a first preset height threshold to obtain first area point cloud data and second area point cloud data, wherein the first area point cloud data is point cloud data of an area corresponding to a height lower than or equal to the first preset height threshold, a first downsampling coefficient is set corresponding to the first area, the second area point cloud data is point cloud data corresponding to a height higher than the first preset height threshold, a second downsampling coefficient is set corresponding to the second area point cloud data, the perception internal downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient;

[0053] For example, in the downsampled boundary area point cloud data, there may be small-target objects such as cones and roadblocks with relatively small heights, and there may also be obstacles such as vehicles with relatively large heights. Small-target objects such as cones and roadblocks are generally short in height, and the point cloud space can continue to be segmented by the first preset height in the above-mentioned P1 sub-point cloud space. Select the first preset height threshold h0 (the value h0 can be 0.5m), and the point clouds greater than or equal to the first preset height constitute the sub-point cloud P3 (second area point cloud data), and the point clouds less than or equal to the first preset threshold constitute the sub-point cloud P4 (first area point cloud data). P4 is mainly used to identify low and small targets, and its downsampling grid can be set to a smaller size to retain more points for subsequent clustering and other processes. The downsampling grid parameters in the P3 and P4 spaces are denoted as a3 and a4, where n1 and n2 are positive integers, n1 is the second downsampling coefficient, and n2 is the first downsampling coefficient. The value range can be defined as n1∈[2,4], n2>n1, is the second downsampling coefficient, is the height value of the first downsampling coefficient z. a0 is the downsampling initial grid parameter, which is used for downsampling the P2 sub-point cloud and can generally be between 0.3cm and 0.5cm.

[0054]

[0055] S130: Perform target recognition on the first area point cloud data and the second area point cloud data to obtain target object information.

[0056] For example, the point cloud data of the first area and the point cloud data of the second area can be merged to obtain the point cloud data for identifying obstacles obtained by the target vehicle at the current position. By using a clustering algorithm or a deep learning algorithm to identify the downsampled point cloud data, obstacles near the target vehicle can be identified quickly and accurately, guiding the target vehicle to perform intelligent driving.

[0057] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application performs point cloud boundary segmentation on the original point cloud data based on the preset downsampling boundary and boundary width to obtain downsampled boundary area point cloud data, and then divides the above-mentioned downsampled boundary area point cloud data into boundary low sub-areas and boundary non-low sub-areas according to the height according to the first preset height threshold, and then calculates the first downsampling coefficient and the second downsampling coefficient using the preset default downsampling parameters. Downsampling operations are performed at different scales in different point cloud spaces, which can effectively reduce the number of points in the point cloud to ensure the efficiency of subsequent processing, and maintain the point cloud features to ensure the accuracy of subsequent clustering and other operations.

[0058] In some examples, the target object information includes object type information of the target object;

[0059] The above method further includes:

[0060] A third downsampling coefficient of the corresponding area of ​​the target object is determined based on the object type information.

[0061] Exemplarily, after the above-mentioned boundary segmentation and height segmentation operations, the type information of the object can be obtained by identifying the point cloud data of the first area and the point cloud data of the second area. The object can be a small fixed object, a large fixed object, a small mobile object, a large mobile object, etc. After preliminary identification of the object, the downsampling coefficient is adjusted according to the type of the target object, and the third downsampling coefficient of the corresponding area of ​​different objects is re-determined to meet the needs of identifying different types of objects and effectively improve the recognition speed.

[0062] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application can further adjust its downsampling coefficient according to the type information of the object after preliminary identification of the object to meet the needs of identifying different types of objects and effectively improve the recognition speed.

[0063] In some examples, determining the third downsampling coefficient for the corresponding area of ​​the target object based on the object type information includes:

[0064] When the height of the target object is less than or equal to the first preset height threshold, the third downsampling coefficient corresponding to the target object is set to the first downsampling coefficient;

[0065] and / or,

[0066] When the height of the target object is greater than the first preset height threshold and the length is greater than the length threshold, or the height of the target object is greater than the first preset height threshold or the width is greater than the width threshold, setting the third downsampling coefficient corresponding to the target object to the perceptual internal downsampling coefficient;

[0067] and / or,

[0068] When the height of the target object is greater than the first preset height threshold and the length is less than or equal to the length threshold, the third downsampling coefficient corresponding to the target object is set as the downsampling coefficient of the point cloud in the non-low area in the downsampling boundary;

[0069] and / or,

[0070] When the height of the target object is greater than the first preset height threshold and the width is less than the width threshold, the third downsampling coefficient corresponding to the target object is set as the downsampling coefficient of the point cloud in the non-low area in the downsampling boundary.

[0071] For example, Figure 3As shown in the figure, the motion trajectory of the perceived target object is tracked and predicted in the P2 sub-point cloud space through historical perception results. The possible position of the target object in the current frame is spatially segmented, and the sub-spaces P5, P6, and P7 of the target range are divided according to the size of the target. The remaining sub-point cloud space of the non-target area is recorded as P8, and the corresponding downsampling grid parameters are a5, a6, a7, and a8. The calculation method of a5, a6, and a7 is related to the size and / or height of the target object, as shown in the following formula:

[0072]

[0073] Where l is the length of the target object, w is the width of the target object, h is the height of the target object, h0, l0, and w0 are the height threshold, length threshold, and width threshold, respectively. The reference values ​​are h0 = 0.5m, l0 = 3m, and w0 = 2m. When h ≤ h0, that is, the target object is less than or equal to the first preset height threshold, the target object is a small target, and the third downsampling coefficient is set to the first downsampling coefficient. If the target object is greater than the second preset height threshold and l > l0 or w > w0, the target object is likely a medium-to-large vehicle-type target, and the original default parameters can be used. Otherwise, the target object may be a pedestrian or rider, and the downsampling coefficient of the non-low area point cloud at the downsampling boundary can be used. P8 is not a key perception area, and the default downsampling parameters can be used for downsampling.

[0074] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application readjusts the third downsampling coefficient of the corresponding area according to the size of the target object, and makes targeted adjustments to the point cloud accuracy of different types of objects, thereby ensuring recognition accuracy while improving the target object recognition speed.

[0075] In some examples, the method further includes:

[0076] In the case where there is an overlapping area between the corresponding areas of the plurality of target objects, a smaller downsampling coefficient is selected among the corresponding areas of the overlapping target objects to perform downsampling processing on the overlapping area.

[0077] For example, when the areas corresponding to multiple target objects overlap, the overlapping areas are downsampled using a smaller downsampling coefficient to ensure that the point cloud density of the overlapping areas is sufficient to identify any one of the overlapping target objects, thereby avoiding a decrease in recognition accuracy due to area overlap.

[0078] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application uses a smaller downsampling coefficient to downsample the overlapping area when there are overlapping areas of multiple target objects, thereby ensuring the recognition accuracy of the target objects.

[0079] In some examples, the method further includes:

[0080] When a smaller target object overlaps behind a larger target object, obtaining a movement pattern of the smaller target object;

[0081] Downsampling processing is performed in the overlapping area based on the movement rule and the smaller downsampling coefficient.

[0082] For example, when a smaller target object overlaps behind a larger target object, the larger object may block the smaller target object for a period of time in the future, and after a period of time, the smaller object will suddenly appear in the area corresponding to the non-target object. At this time, the downsampling coefficient of the area corresponding to the non-target object is larger, and the point cloud data is indeed more, which affects the recognition of the smaller object, and the speed and accuracy will be reduced. In order to avoid this situation, when the smaller target object overlaps behind the larger target object, the movement pattern of the smaller target object is first obtained, and the overlapping area is downsampled according to the movement pattern of the smaller object and the smaller downsampling coefficient. For example, if the smaller object moves at a constant speed from left to right, then a moving overlapping area is set according to the movement pattern and the smaller downsampling coefficient. At this time, if the smaller object suddenly appears in the point cloud data obtained by the vehicle, the downsampling coefficient of the area corresponding to the smaller object is also sufficient to meet the requirements for identifying the object.

[0083] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application obtains the movement pattern of the smaller target object when a smaller target object overlaps behind a larger target object, and processes it. This can ensure that when the smaller object jumps out of the occlusion of the larger object, the downsampling coefficient of the area corresponding to the smaller object is sufficient to meet the requirements for identifying the object, thereby avoiding missed identification and causing danger.

[0084] In some examples, the method further includes:

[0085] In the case where the target object is a moving object, obtaining a moving speed of the moving object;

[0086] A fourth downsampling coefficient of the area corresponding to the moving object is determined based on the moving speed.

[0087] For example, when the target object is a moving object, the moving speed of the moving object is obtained and the fourth downsampling coefficient is determined according to the moving speed of the object. The faster the moving speed of the object, the smaller the fourth downsampling coefficient can be set. The faster the moving speed of the target object, the greater the potential risk to the vehicle. When the vehicle speed is higher, the fourth downsampling coefficient can be set to be smaller, that is, more point cloud information can be retained to make better and more accurate identification of its motion patterns and the characteristics of the object, thereby improving the safety of the vehicle's autonomous driving.

[0088] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application dynamically adjusts the downsampling coefficient according to the moving speed of the target object, which can effectively make more accurate identification of moving objects at different speeds.

[0089] In some examples, the method further includes:

[0090] In the case where the target object is a moving object, obtaining a moving direction of the moving object;

[0091] In a case where the moving direction of the object does not overlap with the moving direction of the target vehicle, the perception internal downsampling coefficient is set as the downsampling coefficient of the area corresponding to the moving object.

[0092] For example, when the target object is identified as a moving object, the moving direction of the moving object can be obtained based on the radar point cloud data of the target vehicle identified in different time periods. If the moving direction of the object does not overlap with the moving direction of the target vehicle, that is, the moving object will not cause a risk of collision with the target vehicle, there is no need to use a larger grid parameter to identify the object. At this time, the downsampling coefficient of the area corresponding to the moving object is set to the perception internal downsampling coefficient (that is, the default downsampling coefficient).

[0093] In summary, the radar point cloud data downsampling processing method provided in the embodiment of the present application determines the moving direction of a moving object. When the moving direction of the object does not interfere with the driving of the vehicle, its grid parameters are reduced to improve the speed of obstacle recognition.

[0094] In some examples, such as Figure 4 As shown, the point cloud space can be divided into several subspaces P = {P1, P2, P3, ...}, and the downsampling method F = {f1, f2, f3, ...} and the parameter A = {a1, a2, a3, ...} of each subspace are set according to the needs. Finally, the downsampled sub-point clouds are merged to obtain the final point cloud downsampling result. The spatial segmentation and parameter setting are adjusted according to specific factors such as vehicle type, operating environment, and identification target. The present invention will provide several specific implementation schemes. The overall flow chart is shown as follows: Figure 4As shown in the figure: the original point cloud is divided into different sub-point clouds through three methods: boundary segmentation, height segmentation and perception tracking segmentation; each sub-point cloud is downsampled by setting its own downsampling method parameters; finally, the downsampling results of each sub-point cloud are merged to obtain the final point cloud downsampling result.

[0095] See also Figure 5 An embodiment of the radar point cloud data downsampling processing device in the embodiment of the present application may include:

[0096] an acquisition unit 21 for acquiring downsampled boundary area point cloud data of the target vehicle, wherein the downsampled boundary area point cloud data is acquired by performing boundary segmentation processing on the original point cloud data acquired from the target vehicle based on a preset downsampled boundary width, a downsampled boundary downsampling coefficient, and a perceived internal downsampling coefficient;

[0097] a segmentation unit 22, configured to perform height segmentation on the downsampled boundary area point cloud data according to a first downsampling coefficient and a second downsampling coefficient based on a first preset height threshold to obtain first area point cloud data and second area point cloud data, wherein the first area point cloud data is point cloud data of an area corresponding to a height lower than or equal to the first preset height threshold, the second area point cloud data is point cloud data corresponding to a height higher than the first preset height threshold, the perception internal downsampling coefficient is greater than the downsampling boundary downsampling coefficient, the downsampling boundary downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient;

[0098] The recognition unit 23 is configured to perform target recognition on the first area point cloud data and the second area point cloud data to obtain target object information.

[0099] like Figure 6 As shown, an embodiment of the present application further provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, the steps of any of the above-mentioned methods for downsampling processing of radar point cloud data are implemented.

[0100] Since the electronic device introduced in this embodiment is a device used to implement a radar point cloud data downsampling processing device in the embodiment of the present application, based on the method introduced in the embodiment of the present application, technical personnel in this field can understand the specific implementation of the electronic device of this embodiment and its various variations. Therefore, how the electronic device implements the method in the embodiment of the present application is no longer introduced in detail here. As long as the equipment used by technical personnel in this field to implement the method in the embodiment of the present application falls within the scope of protection to be protected by this application.

[0101] In a specific implementation process, the computer program 311 can be implemented when executed by a processor. Figure 1 Any implementation manner in the corresponding embodiments.

[0102] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0103] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0104] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0105] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0107] The present application also provides a computer program product, which includes computer software instructions. When the computer software instructions are executed on a processing device, the processing device is caused to execute the following Figure 1 The process of downsampling the radar point cloud data in the corresponding embodiment.

[0108] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state drive (SSD)).

[0109] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0111] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0112] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0113] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0114] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A radar point cloud data downsampling processing method, characterized in that: include: The original point cloud data obtained from the target vehicle is subjected to boundary segmentation processing based on the preset downsampling boundary and boundary width to obtain an outer boundary area, a downsampling boundary area, and an inner area, wherein the downsampling boundary area is downsampled using the downsampling boundary downsampling coefficient; the inner area is downsampled using the perceived inner downsampling coefficient; Acquire downsampled boundary area point cloud data of the downsampled boundary area of ​​the target vehicle; The downsampled boundary area point cloud data is highly segmented according to a first preset height threshold, and is divided into a boundary low area and a boundary non-low area to obtain first area point cloud data and second area point cloud data, wherein the boundary low area corresponds to the first area point cloud data, which is point cloud data with a height lower than or equal to the first preset height threshold, and the boundary non-low area corresponds to the second area point cloud data, which is point cloud data with a height higher than the first preset height threshold, a first downsampling coefficient is set corresponding to the first area point cloud data, and a second downsampling coefficient is set corresponding to the second area point cloud data, the perception internal downsampling coefficient is greater than the downsampling boundary downsampling coefficient, the downsampling boundary downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient; Performing target recognition on the first area point cloud data and the second area point cloud data to obtain target object information; Based on the target object information, a third downsampling coefficient of a corresponding area of ​​the target object is determined.

2. The method according to claim 1, wherein The target object information includes object type information of the target object; The method further comprises: Based on the object type information, a third downsampling coefficient of a corresponding area of ​​the target object is determined.

3. The method according to claim 2, wherein The determining, based on the object type information, a third downsampling coefficient of a corresponding area of ​​the target object includes: When the height of the target object is less than or equal to the first preset height threshold, setting the third downsampling coefficient corresponding to the target object to the first downsampling coefficient; and / or, When the height of the target object is greater than the first preset height threshold and the length is greater than the length threshold, or the height of the target object is greater than the first preset height threshold, or the height is greater than the first preset height threshold and the width is greater than the width threshold, setting the third downsampling coefficient corresponding to the target object to the perception internal downsampling coefficient; and / or, When the height of the target object is greater than the first preset height threshold and the length is less than or equal to the length threshold, the third downsampling coefficient corresponding to the target object is set as the downsampling coefficient of the point cloud in the non-low area in the downsampling boundary; and / or, When the height of the target object is greater than the first preset height threshold and the width is less than the width threshold, the third downsampling coefficient corresponding to the target object is set as the downsampling coefficient of the point cloud in the non-low area in the downsampling boundary.

4. The method according to claim 2, wherein Also includes: In the case where there is an overlapping area between the corresponding areas of the plurality of target objects, a smaller downsampling coefficient is selected among the corresponding areas of the overlapping target objects to perform downsampling processing on the overlapping area.

5. The method according to claim 4, wherein Also includes: When a smaller target object overlaps behind a larger target object, obtaining a movement pattern of the smaller target object; Downsampling processing is performed in the overlapping area based on the movement rule and the smaller downsampling coefficient.

6. The method according to claim 2, wherein Also includes: When the target object is a moving object, obtaining a moving speed of the moving object; Based on the moving speed, a fourth downsampling coefficient of the area corresponding to the moving object is determined.

7. The method according to claim 2, wherein Also includes: When the target object is a moving object, obtaining a moving direction of the moving object; In a case where the moving direction of the object does not overlap with the moving direction of the target vehicle, the perception internal downsampling coefficient is set as the downsampling coefficient of the area corresponding to the moving object.

8. A radar point cloud data downsampling processing device, characterized in that: include: an acquisition unit, configured to perform boundary segmentation processing on the original point cloud data acquired from the target vehicle based on a preset downsampling boundary and a boundary width to obtain an outer boundary area, a downsampling boundary area, and an inner area, wherein the downsampling boundary area is downsampled using a downsampling boundary downsampling coefficient; the inner area is downsampled using a perceived inner downsampling coefficient; and obtain downsampling boundary area point cloud data of the downsampling boundary area of ​​the target vehicle; A segmentation unit is configured to perform height segmentation on the downsampled boundary area point cloud data according to a first preset height threshold, dividing the downsampled boundary area point cloud data into a boundary low area and a boundary non-low area, so as to obtain first area point cloud data and second area point cloud data, wherein the boundary low area corresponds to the first area point cloud data, which is point cloud data with a height lower than or equal to the first preset height threshold, the boundary non-low area corresponds to the second area point cloud data, which is point cloud data with a height higher than the first preset height threshold, a first downsampling coefficient is set corresponding to the first area point cloud data, a second downsampling coefficient is set corresponding to the second area point cloud data, the perception internal downsampling coefficient is greater than the downsampling boundary downsampling coefficient, the downsampling boundary downsampling coefficient is greater than the second downsampling coefficient, and the second downsampling coefficient is greater than the first downsampling coefficient; an identification unit, configured to perform target identification on the first area point cloud data and the second area point cloud data to obtain target object information; Based on the target object information, a third downsampling coefficient of a corresponding area of ​​the target object is determined.

9. An electronic device comprising: A memory and a processor, wherein the processor is configured to implement the steps of the radar point cloud data downsampling processing method according to any one of claims 1 to 7 when executing a computer program stored in the memory.

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

Images