Drivable Area Detection Method, Device, and Storage Medium

Abstract

The present application discloses a method, device, and storage medium for detecting a drivable area. The method includes: determining a target road boundary from several road boundaries in a detection area; partitioning the detection area based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary, where the probability that the first type of partition is within the target road boundary meets the requirements; generating the drivable area in the first type of partition using a first area generation method that takes less time, generating the drivable area in the second type of partition using a second area generation method, and using the first area generation method to generate the drivable area based on the road boundary passing through the first type of partition; fusing the drivable areas in the first type of partition and the second type of partition to obtain the drivable area of the target vehicle in the detection area. By the above method, the present application can improve the detection efficiency while ensuring the detection effect of the drivable area.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular, to a method, device, and storage medium for detecting a drivable area. Background Art

[0002] The drivable area of a vehicle, also known as the reachable space of the vehicle, indicates the area that the vehicle can plan and control. Determining the drivable area of the vehicle has a very important reference role in planning and controlling the driving path of the vehicle, and plays a crucial "backup" role in the driving safety of assisted driving or autonomous driving.

[0003] Therefore, how to accurately and quickly determine the drivable area of the vehicle is very important for path planning and control. Summary of the Invention

[0004] The main technical problem to be solved by this application is to provide a method, device, and storage medium for detecting a drivable area, which can improve the detection efficiency while ensuring the detection effect of the drivable area.

[0005] To solve the above technical problem, a technical solution adopted by this application is: providing a method for detecting a drivable area, the method includes: determining a target road boundary from several road boundaries in a detection area, the detection area is within a preset distance range of a target vehicle; partitioning the detection area based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary, the probability that the first type of partition corresponding to the target road boundary is within the target road boundary meets the requirement; generating a drivable area in the first type of partition by using a first area generation method, and generating a drivable area in the second type of partition by using a second area generation method, the processing time of the first area generation method is less than that of the second area generation method, and the first area generation method is based on the road boundary passing through the first type of partition to generate a drivable area; fusing the drivable area in the first type of partition and the drivable area in the second type of partition to obtain the drivable area of the target vehicle in the detection area.

[0006] Among them, partitioning the detection area based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary includes: finding a target position point in the target road boundary; using the connection line between the target position point and the position where the target vehicle is located as a partition line; using the partition line corresponding to the target road boundary to divide at least part of the detection area into a first type of partition and a second type of partition corresponding to the target road boundary, the first type of partition corresponding to the target road boundary is on the side of the partition line far from the target road boundary, and the second type of partition is on the side of the partition line close to the target road boundary.

[0007] Wherein, the target position point is the farthest position point in the target road boundary and along the preset extension direction of the target road boundary, and the preset extension direction corresponds to the advancing direction of the target vehicle; and / or, using the dividing line corresponding to the target road boundary, at least part of the detection area is divided into a first type of partition and a second type of partition corresponding to the target road boundary, including: for each target road boundary, the area between the reference line of the target vehicle and the dividing line corresponding to the target position point of the target road boundary is used as the first type of partition corresponding to the target road boundary, and the area on the side of the dividing line close to the target road boundary in the detection area is determined as the second type of partition corresponding to the target road boundary, and the reference line passes through the target vehicle and extends along the length direction of the road where the target vehicle is located.

[0008] Wherein, before determining the target road boundary from several road boundaries in the detection area, the method further includes: obtaining road boundary data of the detection area, where the road boundary data includes the boundary position information of the road boundary and the confidence level regarding the boundary position information; determining the target road boundary from several road boundaries in the detection area, including: determining the road boundary whose confidence level meets the preset confidence level requirement as the target road boundary.

[0009] Wherein, the preset confidence level requirement is that the confidence level is less than or equal to the first confidence level threshold and greater than or equal to the second confidence level threshold; before generating the drivable area in the first type of partition using the first area generation method and generating the drivable area in the second type of partition using the second area generation method, it further includes: obtaining the road boundary with a confidence level greater than the first confidence level threshold as the first road boundary, and determining the area on the side of the target vehicle close to the first road boundary as the first type of partition; and / or, obtaining the road boundary with a confidence level less than the second confidence level threshold as the second road boundary, and determining the area on the side of the target vehicle close to the second road boundary as the second type of partition.

[0010] Wherein, determining the area on the side of the target vehicle close to the first road boundary as the first type of partition includes: determining the area on the first target side of the reference line of the target vehicle in the detection area as the first type of partition, where the first target side is the side close to the first road boundary, and the reference line passes through the target vehicle and extends along the length direction of the road where the target vehicle is located; and / or, determining the area on the side of the target vehicle close to the second road boundary as the second type of partition includes: determining the area on the second target side of the reference line of the target vehicle in the detection area as the second type of partition, where the second target side is the side close to the second road boundary.

[0011] Among them, generating the drivable area in the first type of partition using the first area generation method includes: performing object detection on the first type of partition using a target processing algorithm to obtain a processing result, where the target processing algorithm includes at least one of an object detection algorithm and an object tracking algorithm, and the processing result includes position representation information of the detection frames of several obstacles; obtaining the drivable area in the first type of partition based on the processing result and the road boundary in the first type of partition; and / or, generating the drivable area in the second type of partition using the second area generation method includes: performing semantic segmentation on the second type of partition using a semantic segmentation point algorithm to obtain point cloud data; obtaining the drivable area in the second type of partition based on the point cloud data.

[0012] Among them, the position representation information of the detection frame of the obstacle includes the central position of the center point of the obstacle and size data; obtaining the drivable area in the first type of partition based on the processing result and the road boundary in the first type of partition includes: for each obstacle, determining the position information of each vertex on the detection frame of the obstacle based on the central position and size data; determining the vertex pairs of each side belonging to the detection frame based on the magnitude relationship between the angle between each connection line of the detection frame and the first preset direction, where each connection line is the connection line between each vertex and the center point; using the position information of the vertex pairs of each side corresponding to each obstacle to select the target side closest to the target vehicle in different second preset directions from each obstacle; determining the candidate drivable area of the target vehicle from the first type of partition based on the target sides corresponding to each obstacle; and selecting the area within the road boundary in the first type of partition from the candidate drivable area of the target vehicle as the drivable area of the first type of partition.

[0013] Among them, obtaining the drivable area in the second type of partition based on the point cloud data includes: extracting the ground point cloud belonging to the ground from the point cloud data; processing the ground point cloud to obtain the outer envelope of the ground point cloud; and determining the drivable area in the second type of partition based on the outer envelope.

[0014] Among them, before determining the target road boundary from several road boundaries in the detection area, it further includes: using the sensing data detected in the detection area to determine multiple road boundary points; and fitting several road boundaries using the multiple road boundary points.

[0015] To solve the above technical problems, another technical solution adopted by this application is: providing an electronic device, including a memory and a processor coupled to each other, where the memory stores program instructions; the processor is used to execute the program instructions stored in the memory to implement the above method.

[0016] To solve the above technical problems, yet another technical solution adopted by this application is: providing a computer-readable storage medium for storing program instructions, and the program instructions can be executed to implement the above method.

[0017] In the above solution, after determining the target road boundary from the detection area, the detection area is partitioned based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary. Then, a drivable area in the first type of partition is generated using a first area generation method, and a drivable area in the second type of partition is generated using a second area generation method. Next, the drivable areas in the first type of partition and the second type of partition are fused to obtain the drivable area of the target vehicle in the detection area. Since the probability that the first type of partition is within the target road boundary meets the requirements, the generation of the drivable area in the first type of partition is not affected by the target road boundary. Therefore, the use of the first area generation method in this application does not affect the detection effect of the drivable area in the first type of partition. Further, since the processing time of the first area generation method is less than the processing time of the second area generation method, compared with the method of generating the vehicle drivable area using the second area generation method for each partition, this application can not only ensure the detection effect of the drivable area in the entire detection area, but also reduce the detection time to improve the detection efficiency of the drivable area. Description of the Drawings

[0018] Figure 1 is a schematic flowchart of an embodiment of the drivable area detection method provided by this application;

[0019] Figure 2 is Figure 1 a schematic flowchart of an embodiment of step S12 shown;

[0020] Figure 3 is a schematic diagram of the area division provided by this application;

[0021] Figure 4 is Figure 1 a partial schematic flowchart of an embodiment of step S13 shown;

[0022] Figure 5 is Figure 4 a schematic flowchart of an embodiment of step S42 shown;

[0023] Figure 6 is a schematic diagram of the obstacle detection frame provided by this application;

[0024] Figure 7 is Figure 1 a partial schematic flowchart of an embodiment of step S13 shown;

[0025] Figure 8 is a schematic framework diagram of an embodiment of the electronic device provided by this application;

[0026] Figure 9 is a schematic framework diagram of the computer-readable storage medium provided by this application. Detailed implementation manners

[0027] To make the objectives, technical solutions and effects of this application clearer and more definite, the following further describes this application in detail with reference to the accompanying drawings and by way of examples.

[0028] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of this application, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be construed as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first", "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between various embodiments may be combined with each other, but it must be based on the fact that those skilled in the art can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0029] Please refer to Figure 1 , Figure 1 , which is a schematic flowchart of an embodiment of the drivable area detection method provided by this application. It should be noted that if there are substantially the same results, this embodiment is not limited to Figure 1 the process sequence shown. As Figure 1 shown, this embodiment includes:

[0030] S11: Determine a target road boundary from several road boundaries in the detection area, where the detection area is within a preset distance range of the target vehicle.

[0031] This embodiment is used to first determine a target road boundary from the detection area, then partition the detection area based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary, and then use different area generation methods to generate the drivable area in the first type of partition and the drivable area in the second type of partition respectively, and fuse the drivable areas of the two types of partitions to obtain the drivable area of the target vehicle in the entire detection area.

[0032] In this embodiment, the detection area is an area within a preset distance range of the target vehicle, where the detection area includes the area where the current road on which the target vehicle is traveling is located. Specifically, the preset distance range can be determined according to the actual situation. Exemplarily, considering the width information of the road, the area within the preset range in front of the target vehicle and the preset ranges on the left and right can be used as the detection area. For example, the area 70 meters in front of the target vehicle and 20 meters on each of the left and right can be used as the detection area; further, different preset distance ranges can also be set according to the driving speed of the target vehicle to make the detection area adapt to the driving speed of the vehicle.

[0033] The road boundaries described in this article correspond to the curbs of the road on which the target vehicle travels in the actual scenario, rather than the lane lines of each lane on the road. Generally, the number of road boundaries of a road is two, but for intersection scenarios, the number of road boundaries is greater than two (for example, 4, 6, or 8, etc.).

[0034] It should be noted that the target road boundaries described in this article can be some of the road boundaries among several road boundaries, or all of the road boundaries among several road boundaries.

[0035] In one implementation, considering that the confidence level of the road boundary will affect the detection of the drivable area, in order to ensure the accuracy and efficiency of the drivable area detection, the target road boundary can be determined first according to the confidence of the road boundary, and then the detection area can be partitioned according to the target road boundary determined by the confidence, into a first type of partition where the probability within the target road boundary meets the requirements, and a second type of partition where it is uncertain whether it is within or outside the target road boundary. Since the probability of the first type of partition within the target road boundary meets the requirements, for the determination of the drivable area of this first type of partition, it is not affected by the confidence of the road boundary, and the drivable area in the first type of partition can be generated using the first area generation method with a shorter processing time; while for the second type of partition where it is uncertain whether it is within or outside the target road boundary, the determination of the drivable area of this second type of partition may be affected by the confidence of the road boundary. Therefore, in order to ensure the detection accuracy of the drivable area of the second type of partition, the drivable area in the second type of partition can be generated using the second area generation method with a longer processing time but better processing effect, and then the drivable areas in the first type of partition and the second type of partition are fused to obtain the drivable area of the target vehicle in the detection area, so as to improve the detection efficiency while ensuring the detection effect of the drivable area.

[0036] In this implementation, considering that the confidence of the road boundary affects the detection of the drivable area, before determining the target road boundary from several road boundaries in the detection area, the road boundary data of the detection area (including the boundary position information of the road boundary and the confidence regarding the boundary position information) can be obtained first; and after determining the confidence of the road boundary regarding the boundary position information, the road boundary with a confidence meeting the preset confidence requirement is determined as the target road boundary, and then step S12 and subsequent steps are executed. Among them, the preset confidence requirement is that the confidence is less than or equal to the first confidence threshold and greater than or equal to the second confidence threshold, that is, select the road boundaries within a certain confidence range (for example, 30%-80%) from several road boundaries as the target road boundaries. The specific preset confidence requirement can be determined according to the actual situation to ensure the detection effect of the drivable area while improving the detection efficiency.

[0037] In this embodiment, the target road boundary determined according to whether the confidence of the road boundary meets the confidence requirement may be one or more boundaries among several road boundaries. Specifically, it can be determined according to the confidence of the detected road boundary position information. Among them, the existing detection method can be used to obtain the road boundary data (including the confidence of the boundary position information) of the detection area. For example, the sensing data detected in the detection area can be first used to determine a plurality of road boundary points, and then a plurality of road boundary points can be used to fit a plurality of road boundaries. Among them, the sensing data can be the data detected by relevant sensors, and the sensor can be, but is not limited to, a lidar, and can also be a millimeter-wave radar or a camera, etc.; each of the fitted road boundaries may include one or more of the parameters such as the coefficients of the road equation, the starting point coordinates, the effective length, and the confidence. The determined target road boundary may be one or more boundaries among several road boundaries. Specifically, it can be determined according to the confidence of the detected road boundary position information.

[0038] In an implementation scenario, if the confidence of the road boundary is greater than the first confidence threshold, it indicates that the credibility of the currently obtained road boundary is relatively high. In this scenario, the determination of the drivable area in the detection area is not affected by the road boundary, or the possibility of being affected by the road boundary is very low. To improve the detection efficiency, the road boundary with a confidence greater than the first confidence threshold can be obtained as the first road boundary, and the area on the side of the target vehicle close to the first road boundary can be determined as the first type of partition, so as to facilitate the subsequent generation of the drivable area in the first type of partition by using the first area generation method, that is, the drivable area of the entire detection area.

[0039] In another implementation scenario, if the confidence of the road boundary is less than the first confidence threshold, it indicates that the credibility of the currently obtained road boundary is relatively low, and this road boundary is very likely not the real road boundary. In this scenario, the probability that the determination of the drivable area in the detection area is affected by the road boundary is relatively large. To ensure the detection accuracy (detection effect), the road boundary with a confidence less than the first confidence threshold can be obtained as the second road boundary, and the area on the side of the target vehicle close to the second road boundary can be determined as the second type of partition, so as to facilitate the subsequent generation of the drivable area in the second type of partition by using the second area generation method, that is, the drivable area of the entire detection area.

[0040] Specifically, a reference line of the target vehicle can be determined in advance, and the first type of partition or the second type of partition can be determined by using the reference line. In a specific embodiment, the area on the first target side of the reference line of the target vehicle in the detection area is determined as the first type of partition, and the first target side is the side close to the first road boundary; the area on the second target side of the reference line of the target vehicle in the detection area is determined as the second type of partition, and the second target side is the side close to the second road boundary. Among them, the reference line passes through the target vehicle and extends along the length direction of the road where the target vehicle is located.

[0041] Please refer to Figure 3 , Figure 3 which is the schematic diagram of area division provided by this application. As Figure 3 shown, the point O represents the position of the target vehicle, OX represents the reference line, the detection area includes all areas from S1 to S6. If the confidence levels of the road boundaries l1 and l2 are both greater than the first confidence threshold, then l1 and l2 are respectively the first road boundaries, and the areas corresponding to S1 - S6 are all the first type of partitions; if the confidence levels of the road boundaries l1 and l2 are both less than the first confidence threshold, then l1 and l2 are respectively the second road boundaries, and the areas corresponding to S1 - S6 are all the second type of partitions. If the confidence level of the road boundary l1 is greater than the first confidence threshold and the confidence level of the road boundary l2 is less than the first confidence threshold, then l1 is the first road boundary and l2 is the second road boundary. In this case, the areas corresponding to S1 - S3 are the first type of partitions on the first target side of the reference line of the target vehicle, and the areas corresponding to S4 - S6 are the second type of partitions on the second target side of the reference line of the target vehicle.

[0042] Of course, in other embodiments, when there are no high requirements for processing time or detection accuracy, the influence of the confidence level of the road boundary on the detection of the drivable area may not be considered, and all road boundaries are regarded as target road boundaries, and then step S12 and subsequent processing are executed.

[0043] S12: Partition the detection area based on the target road boundary to obtain the first type of partition and the second type of partition corresponding to the target road boundary, and the probability that the first type of partition corresponding to the target road boundary is within the target road boundary meets the requirements.

[0044] This embodiment mainly partitions the detection area into the first type of partition and the second type of partition. Among them, the probability that the first type of partition corresponding to the target road boundary obtained by partitioning is within the target road boundary meets the requirements, that is, the probability that the area corresponding to the first type of partition in this article is within the real road boundary of the road where the target vehicle is located meets the requirements. For the specific reasons for partitioning in this embodiment, please refer to the relevant description in step S11 above.

[0045] In this embodiment, in order to distinguish the above-mentioned first type of partition and the second type of partition, an appropriate partition line can be determined according to the target road boundary, and the detection area is divided into the first type of partition and the second type of partition corresponding to the target road boundary through the partition line.

[0046] Specifically, please refer to Figure 2 , Figure 2 which Figure 1 is a schematic flowchart of an embodiment of step S12 shown in. It should be noted that if there are substantially the same results, this embodiment is not limited to Figure 2 the Figure 2 flow order shown in. As

[0047] S21: Find the target position points in the target road boundary.

[0048] In this embodiment, the target position points are first found from the target road boundary, and then the partition line is determined according to the target position points and the position where the target vehicle is located. Furthermore, the first type of partition and the second type of partition corresponding to the target road boundary are obtained by dividing with the partition line.

[0049] It should be noted that the target position points are the position points in the target road boundary, and the farther the target position points are from the target vehicle, the higher the probability that the area corresponding to the finally divided first type of partition is within the target road boundary. Therefore, in one implementation manner, in order to better ensure the detection effect, the farthest position point in the target road boundary and in the preset extension direction of the target road boundary can be used as the target position point. Among them, the preset extension direction corresponds to the advancing direction of the target vehicle. It should be noted that the preset extension direction is the direction corresponding to the advancing direction of the target vehicle on the target road boundary, rather than the direction opposite to the advancing direction of the target vehicle. It can be understood that the farthest position point in the target road boundary and in the preset extension direction of the target road boundary is the end point in the preset extension direction of the target road boundary.

[0050] Of course, in other implementation manners, the target position points can also be determined according to the confidence level of the boundary position information corresponding to the target road boundary.

[0051] In one embodiment, when the confidence level of the boundary position information corresponding to the target road boundary meets the preset confidence level requirement and the confidence level is relatively high, any position point between the midpoint in the target road boundary and the farthest position point in the preset extension direction of the target road boundary can be used as the target position point, so as to be able to determine the partition line according to the target position point and the position where the target vehicle is located, and further obtain the first type of partition and the second type of partition corresponding to the target road boundary by dividing with the partition line.

[0052] In another embodiment, when the confidence level of the boundary position information corresponding to the target road boundary meets the preset confidence level requirement and the confidence level is relatively low, the farthest position point in the target road boundary and in the preset extension direction of the target road boundary can be used as the target position point, and subsequent operations of determining the partition line and using the partition line to divide the first type of partition and the second type of partition are performed.

[0053] Exemplarily, taking the preset confidence level requirement as 30%-80% as an example, when the confidence level of the boundary position information corresponding to the target road boundary meets 30%-80% and the confidence level is close to 80%, any position point between the midpoint of the target road boundary and the farthest position point in the preset extension direction of the target road boundary can be used as the target position point, and subsequent operations of determining the partition line and using the partition line to divide the first type of partition and the second type of partition are performed; if the confidence level of the boundary position information corresponding to the target road boundary meets 30%-80% but the confidence level is close to 30%, the farthest position point in the target road boundary and in the preset extension direction of the target road boundary can be used as the target position point, and subsequent operations of determining the partition line and using the partition line to divide the first type of partition and the second type of partition are performed.

[0054] S22: Use the line connecting the target position point and the position where the target vehicle is located as the partition line.

[0055] S23: Use the partition line corresponding to the target road boundary to divide at least part of the detection area into the first type of partition and the second type of partition corresponding to the target road boundary.

[0056] Among them, the first type of partition corresponding to the target road boundary is located on the side of the partition line away from the target road boundary, and the second type of partition is located on the side of the partition line close to the target road boundary.

[0057] In one embodiment, at least one target road boundary is determined. For each target road boundary, the area between the reference line of the target vehicle and the partition line corresponding to the target position point of the target road boundary is used as the first type of partition corresponding to the target road boundary, and the area in the detection area on the side of the partition line close to the target road boundary is determined as the second type of partition corresponding to the target road boundary. Among them, the reference line passes through the target vehicle and extends along the length direction of the road where the target vehicle is located.

[0058] Please continue to refer to Figure 3 , such as Figure 3As shown in the figure, the line corresponding to the x-axis is the reference line of the target vehicle. If l1 and l2 are respectively the determined target road boundaries, and for the target road boundary l1, the partition line is the line connecting point A (representing the target position point) and point O (representing the position where the target vehicle is located), then the area where S3 is located is the first type of partition, and the areas where S1 and S2 are located are the second type of partition; similarly, for the target road boundary l2, the partition line is the line connecting point B (representing the target position point) and point O (representing the position where the target vehicle is located), the area where S4 is located is the first type of partition, and the areas where S5 and S6 are located are the second type of partition.

[0059] S13: Generate the drivable area in the first type of partition using the first area generation method, and generate the drivable area in the second type of partition using the second area generation method. The processing time consumption of the first area generation method is less than that of the second area generation method. The first area generation method is based on the road boundary passing through the first type of partition to generate the drivable area.

[0060] In this embodiment, the generation methods used to generate the drivable areas in the first type of partition and the second type of partition are different. Among them, the drivable area in the first type of partition is generated using the first area generation method, and the drivable area in the second type of partition is generated using the second area generation method.

[0061] As can be seen from the above, the probability that the first type of partition corresponding to the target road boundary is within the target road boundary meets the requirements, that is, the probability that the area corresponding to the first type of partition is within the area of the true road boundary in the road where the target vehicle is located meets the requirements, that is, the generation of the drivable area in the first type of partition is not affected or is little affected by the target road boundary. However, it is uncertain whether the area corresponding to the second type of partition is within or outside the target road boundary. Therefore, for the drivable area of the first type of partition that is not affected by the road boundary confidence, the first area generation method with a shorter processing time consumption can be used to generate the drivable area in the first type of partition; for the second type of partition where it is uncertain whether it is within or outside the target road boundary, the determination of the drivable area of this second type of partition may be affected by the road boundary confidence. Therefore, in order to ensure the detection accuracy of the drivable area of the second type of partition, the second area generation method with a longer processing time but better processing effect can be used to generate the drivable area in the second type of partition.

[0062] Among them, the first region generation method is to generate a drivable region based on the road boundary after the first type of partitioning. Optionally, in one embodiment, generating the drivable region in the first type of partition using the first region generation method includes: first performing object detection on the first type of partition using a target processing algorithm to obtain a processing result, and then obtaining the drivable region in the first type of partition based on the processing result and the road boundary in the first type of partition. Among them, the target processing algorithm includes at least one of an object detection algorithm and an object tracking algorithm, and the processing result includes the position representation information of the detection frames of several obstacles. That is, first obtain the position representation information of the detection frames of several obstacles, and then obtain the drivable region in the first type of partition based on the position representation information of the detection frames of several obstacles and the road boundary in the first type of partition. The specific method for generating the drivable region in the first type of partition using the first region generation method is described below Figure 4 corresponding elaboration

[0063] Of course, in other embodiments, the clustering method can also be used to obtain the position representation information of each obstacle, and then obtain the drivable region in the first type of partition based on the position representation information of each obstacle and the road boundary in the first type of partition.

[0064] In another embodiment, generating the drivable region in the second type of partition using the second region generation method includes: first performing semantic segmentation on the second type of partition using a semantic segmentation point algorithm to obtain the point cloud data corresponding to the second type of partition; then obtaining the drivable region in the second type of partition based on the point cloud data. The specific method for generating the drivable region in the second type of partition using the second region generation method is described below Figure 7 corresponding elaboration

[0065] It should be noted that the target processing algorithm used above obtains the position representation information of the detection frames of several obstacles, and the drivable region in the first type of partition is obtained based on the position representation information of the detection frames and the road boundary in the first type of partition; while the point cloud data corresponding to the second type of partition obtained by using the semantic segmentation point algorithm, and the drivable region in the second type of partition is obtained based on the point cloud data. Compared with the second region generation method for region generation based on point cloud data, the first region generation method does not need to use the information of each point in the detection frame for region generation. Therefore, the processing time of the first region generation method is less than that of the second region generation method. However, since the second region generation method takes into account each point cloud data, more data is considered than the first region generation method, and the corresponding detection result will be more accurate.

[0066] S14: Fuse the drivable regions in the first type of partition and the drivable regions in the second type of partition to obtain the drivable region of the target vehicle in the detection region.

[0067] In this embodiment, the drivable areas in the first type of partition and the drivable areas in the second type of partition are obtained, and by combining the drivable areas in the first type of partition and the drivable areas in the second type of partition, the drivable area of the target vehicle in the detection area can be obtained.

[0068] In the above solution, after determining the target road boundary from the detection area, the detection area is partitioned based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary. Then, the drivable area in the first type of partition is generated using the first area generation method, and the drivable area in the second type of partition is generated using the second area generation method. Next, the drivable areas in the first type of partition and the second type of partition are fused to obtain the drivable area of the target vehicle in the detection area. Since the probability that the first type of partition is located within the target road boundary meets the requirements, the generation of the drivable area in the first type of partition is not affected by the target road boundary. Therefore, using the first area generation method does not affect the detection effect of the drivable area in the first type of partition. Further, since the processing time of the first area generation method is less than the processing time of the second area generation method, compared with the method of generating the vehicle drivable area using the second area generation method for each partition, the above method can not only ensure the detection effect of the drivable area in the entire detection area, but also reduce the detection time to improve the detection efficiency of the drivable area.

[0069] Please refer to Figure 4 , Figure 4 For Figure 1 a partial flow schematic diagram of an embodiment of step S13 shown. It should be noted that if there are substantially the same results, this embodiment is not limited to Figure 4 the flow sequence shown. As Figure 4 shown, generating the drivable area in the first type of partition using the first area generation method includes:

[0070] S41: Performing target detection on the first type of partition using a target processing algorithm to obtain a processing result.

[0071] In this embodiment, the target processing algorithm includes at least one of a target detection algorithm and a target tracking algorithm, and the processing result includes the position representation information of the detection frames of several obstacles, so as to represent the position information of the detection frames of the obstacles. Of course, in some embodiments, a relevant clustering algorithm can also be combined to perform target detection on the first type of partition to obtain a processing result. Among them, an obstacle is an object that affects the passage of the target vehicle, which can be a person, a vehicle, a building, or a curb, etc.

[0072] In one embodiment, the position characterization information of the detection frame of the obstacle includes the central position of the center point of the obstacle and relevant dimension data. Among them, the dimension data may include, but is not limited to, one or more of the length, width, and height of the obstacle detection frame, and may also include the orientation angle of the obstacle detection frame. The orientation angle is the included angle between the center line of the detection frame and the preset direction. Among them, the extension direction of the road may be used as the preset direction.

[0073] S42: Based on the processing result and the road boundary in the first type of partition, obtain the drivable area in the first type of partition.

[0074] In this embodiment, after determining the processing result regarding the obstacle in the first type of partition, according to the processing result and the road boundary in the first type of partition, obtain the drivable area in the first type of partition. Among them, the road boundary in the first type of partition represents the road boundary located in the area corresponding to the first type of partition.

[0075] It should be noted that, in some embodiments, the road boundary is the road boundary obtained by detecting the road curb in the detection area, and the target road boundary is the road boundary whose confidence level meets the confidence level requirement among the detected road boundaries; of course, in other embodiments, the detected road boundary is the target road boundary.

[0076] Please refer to Figure 5 , Figure 5 is Figure 4 a schematic flowchart of an embodiment of step S42 shown. It should be noted that, if there are substantially the same results, this embodiment is not limited to Figure 5 the flowchart sequence shown. As Figure 5 shown, based on the processing result and the road boundary in the first type of partition, obtaining the drivable area in the first type of partition includes:

[0077] S51: For each obstacle, based on the central position and dimension data of the obstacle, determine the position information of each vertex on the detection frame of the obstacle.

[0078] In one implementation manner, the dimension data includes the length, width, height, and orientation angle of the obstacle detection frame. The orientation angle is the included angle between the center line of the detection frame and the preset direction. Among them, the preset direction can be determined according to the actual scene. For example, the extension direction of the road can be used as the preset direction. Please refer to Figure 6 , Figure 6 is a schematic diagram of the obstacle detection frame provided by this application. As Figure 6 shown, the positive direction of the X-axis is the preset direction, and θ represents the orientation angle.

[0079] In this embodiment, after determining the central position of the obstacle and the above-mentioned dimension data, the position information of each vertex of the detection frame can be determined by using the following formula:

[0080]

[0081] Wherein, (x A , y A ), (x B , y B ), (x C , y C ) and (x D , y D ) are the coordinates of each vertex of the detection box respectively, θ represents the orientation angle, l, w, and h represent the length, width, and height of the detection box respectively, and (x0, y0) is the coordinate of the center point of the detection box.

[0082] S52: Based on the magnitude relationship between the angles formed by each connection line of the detection box and the first preset direction, determine the vertex pairs of each side belonging to the detection box, where each connection line is the connection line between each vertex and the center point of the obstacle.

[0083] After determining the coordinates (position information) of each vertex of the detection box, use the connection lines between each vertex and the center point of the obstacle as each connection line, and then calculate the angle between each connection line and the first preset direction. For example, Figure 6 as shown, the positive direction of the X-axis is the first preset direction, where the counterclockwise direction is positive. After determining each angle, sort the 4 vertices of the detection box in the counterclockwise direction according to the magnitudes of the angles to determine the sorting result of the 4 vertices of the detection box, and further determine the vertex pairs of each side of the detection box (where the two adjacent vertices in the sorting result are the vertex pairs of each side). In this embodiment, the method of determining the vertex pairs of each side belonging to the detection box through the magnitude relationship of the angles can prevent two vertices that do not belong to the same side from being misused as the vertex pairs on the detection box, resulting in the detected side not being the side corresponding to the adjacent vertices (vertex pairs) of the detection box.

[0084] Please continue to refer to Figure 6 , Figure 6 where the 4 vertices of the detection box are A, B, C, and D respectively, O' is the center point of the detection box, use the connection lines between each vertex and the center point as each connection line, calculate the angles between each connection line and the preset direction, and determine the vertex pairs of each side of the detection box according to the magnitude relationship of the angles. As Figure 6 shown, through the magnitude relationship of the angles, the sorting result of the 4 vertices of the detection box can be determined as ABCD, that is, it is determined that the 4 sides of the detection box are AB, BC, CD, and DA respectively, and AC and BD will not be misused as the sides of the detection box.

[0085] S53: Use the position information of the vertex pairs of each side corresponding to each obstacle to select the target side that is closest to the target vehicle in different second preset directions from each obstacle.

[0086] It can be understood that after determining the position information of the vertex pairs of each side corresponding to each obstacle, the position information of the points between the vertex pairs on each side can be determined using the position information of the vertex pairs of each side, and then the target edges that are closest to the target vehicle in different second preset directions are selected from each obstacle. Herein, the second preset direction is the direction corresponding to the angle between the detection line and the road extension direction.

[0087] Exemplarily, a central point is selected from the target vehicle, and then, starting from this central point, the detection area is divided into several parts using a plurality of detection lines, and the end position information of each detection line on the detection area is recorded. Herein, the end position information includes at least one of the coordinates of the end point of the detection line, the angle between the detection line and the second preset direction, and the length of the detection line. The central point can be selected according to actual needs. For example, the center of the rear axle of the target vehicle can be used as the central point.

[0088] After dividing the detection area into several parts using a plurality of detection lines, each side of each obstacle is traversed using the detection lines to select the target edges that are closest to the target vehicle in different second preset directions from each obstacle.

[0089] Specifically, during the process of traversing each side of the obstacle, the detection lines all start from the central point to detect each obstacle outward. Therefore, the obstacle closest to the target vehicle will be detected first. To select the target edges that are closest to the target vehicle in different second preset directions from each obstacle, for each detection line, the position information of the intersection point of the detection line that first intersects each side of the obstacle is recorded and used as the end point of this detection line, and the position information of the intersection point is used as the end position information of the detection line end point, so as to determine the target edges that are closest to the target vehicle in different second preset directions from each obstacle based on the determined end position information of the detection line.

[0090] Exemplarily, for one side CD of the obstacle detection frame, first determine the angles θ C 、θ D between the connection lines (CO and DO) of each point on CD and the central point O and the road extension direction respectively, and then calculate the detection lines among the plurality of detection lines whose angles are between θ C and θ D . Respectively determine the intersection points of the detection lines whose angles are between θ C and θ D with the CD side. Each intersection point corresponds to the end point of the detection line.

[0091] Exemplarily, the following formula can be used to calculate the intersection point of the detection line and the CD side, that is, the end point coordinates of the detection line:

[0092] Assume that the coordinates of point C are (x C , y C ), the coordinates of point D are (xD , y D ), k θ is the slope of the detection line, and k CD is the slope of side CD, and (x, y) is the end coordinate of the detection line.

[0093] If x C ≠ x D , k θ ≠ k CD , then the end coordinate of the detection line is:

[0094]

[0095] If x C = x D , then the end coordinate of the detection line is:

[0096]

[0097] If x C ≠ x D , or x C ≠ x D , then the end coordinate of the detection line is:

[0098]

[0099] If

[0100] x C ≠ x D , k θ = k CD or x C = x D , or x C =

[0101] x D , then

[0102]

[0103] As can be seen from the above, after determining the intersection points of the detection line and each side, the target side on the obstacle closest to the target vehicle is determined.

[0104] S54: Based on the target sides corresponding to each obstacle, determine the candidate drivable area of the target vehicle from the first type of partition.

[0105] Among them, after determining the target edge corresponding to the obstacle, considering that although there is a distance between different obstacles, there may be a situation where the distance is too close for the target vehicle to pass between the two obstacles. Therefore, to avoid the situation where the target vehicle cannot pass due to insufficient distance between the two obstacles, after determining the target edge corresponding to each obstacle, the distance between the target edges can be determined. If the distance does not meet the distance requirement (the target vehicle cannot pass), the area between the two target edges is determined as an impassable area. Then, in the area where the target vehicle is away from the obstacle corresponding to the target edge, the area other than the determined impassable area is used as the candidate drivable area of the target vehicle, and then the subsequent screening of the drivable area is carried out. For example, first, according to the target edge, the area where the target vehicle is away from the obstacle corresponding to the target edge is determined as A, and then according to the distance between each target edge, the impassable area B in A is determined, and the remaining area A - B is the candidate drivable area of the target vehicle determined from the first type of partition.

[0106] S55: Select the area within the road boundary in the first type of partition from the candidate drivable area of the target vehicle as the drivable area of the first type of partition.

[0107] After determining the candidate drivable area by traversing the obstacles, the area within the road boundary in the first type of partition is used as the drivable area of the first type of partition by using the road boundary for area screening. Among them, to ensure the safe driving of the subsequent target vehicle, after determining the drivable area, small areas where the vehicle cannot pass can be removed according to the area sizes of different areas in the drivable area, and the area after removal is used as the new drivable area. For example, after selecting the area within the road boundary in the first type of partition from the candidate drivable area of the target vehicle, some areas may be relatively close to the road boundary and are not conducive to the passage of the target vehicle, so this part of the area can be removed to update the drivable area of the first type of partition.

[0108] Among them, Figure 5 The method shown is to first determine the candidate drivable area through each obstacle, and then determine the drivable area within the road boundary through the road boundary. Of course, in other embodiments, after determining the position representation information of the detection frames of each obstacle by performing target detection on the first type of partition using the target processing algorithm, the area within the road boundary can be determined by traversing the road boundary first, and then the detection frames of the obstacles within the area within the road boundary are traversed to finally determine the drivable area within the road boundary, or the road boundary can be regarded as an obstacle, and according to the distance information between each obstacle and the target vehicle, the target obstacle with the closest distance to the target vehicle in any direction is selected, and the drivable area within the road boundary is determined according to the target obstacle.

[0109] Please refer to Figure 7, Figure 7 is Figure 1 a partial flowchart of an embodiment of step S13 shown. It should be noted that if there are substantially the same results, this embodiment is not limited to Figure 7 the flowchart order shown. As Figure 7 shown, generating the drivable area in the second type of partition using the second area generation method includes:

[0110] S71: Performing semantic segmentation on the second type of partition using the semantic segmentation point algorithm to obtain point cloud data.

[0111] Among them, the obtained point cloud data includes point clouds belonging to various types of vehicles, road boundary point clouds, point clouds belonging to pedestrians, etc. That is, the point cloud data obtained by semantic segmentation includes point clouds and the types of point clouds. Among them, the types of point clouds are mainly divided into ground point clouds belonging to the ground and obstacle point clouds belonging to obstacles.

[0112] S72: Based on the point cloud data, obtain the drivable area in the second type of partition.

[0113] After obtaining the point cloud data after semantic segmentation, first extract the ground point clouds belonging to the ground (such as sidewalk points, lane line points, points within several road boundaries) from the point cloud data, and then process the ground point clouds to obtain the outer envelope of the ground point clouds, and further determine the drivable area in the second type of partition based on the outer envelope.

[0114] Among them, the area corresponding to the outer envelope is the area within the road boundary. In one embodiment, processing the ground point clouds to obtain the outer envelope of the ground point clouds includes: first obtaining the range of the ground point clouds in the horizontal direction through sorting the ground point clouds; for example, by the X coordinate and Y coordinate of the ground point clouds, determining the X coordinate range and Y coordinate range, evenly dividing the range from the X minimum value to the X coordinate maximum value into N segments along the X-axis direction (usually, the larger N is, the finer the outer envelope is), traversing all the ground point clouds, and according to the ratio of the difference between the X coordinate of the ground point and the X minimum value to the X coordinate range, falling the points onto each segment. For example, for the k-th segment, then traversing to obtain the minimum and maximum Y coordinate values of each segment, and the envelope points (x k , y min ) and (x k , y max ) can be obtained; similarly, evenly dividing the range from the Y coordinate minimum value to the Y coordinate maximum value into N segments along the Y-axis direction, traversing all the ground point clouds, and according to the ratio of the difference between the Y coordinate of the ground point and the Y minimum value to the Y coordinate range, falling the points onto each segment. For the k-th segment, traversing to obtain the minimum and maximum X coordinate values of each segment, and the envelope points (x min , y k ) and (x max , y k),Finally, an outer envelope point set of the ground point cloud is obtained, and then the outer envelope of the ground point cloud is determined through the outer included point set. For example, the outer envelope of the ground point cloud is obtained by fitting the outer included point set.

[0115] After obtaining the outer envelope of the ground point cloud, in the same way as traversing each obstacle along the above detection line, traverse the point cloud of the obstacles in the outer envelope area to select the point cloud closest to the target vehicle in different second preset directions from each obstacle, and then use these point clouds to determine the drivable area in the second type of partition.

[0116] In a specific embodiment, after obtaining the outer envelope point set, first traverse the outer envelope point set in the way of the above detection line. Connect each envelope point in the outer envelope point set with the center point O to determine the included angle between this connection line and the road extension direction, and find the detection line closest to this envelope point according to this included angle. If the detection line has not been detected, determine this envelope point as the end point of the detection line, and record the type corresponding to the end point of this detection line as the ground point cloud, as well as record the length of the current detection line. If this detection line has been detected and the corresponding end point type is the ground point cloud, then compare the length of the detection line closest to this envelope point with the length recorded by this detection line; if it is greater than the length of the recorded detection line, determine this envelope point as the new end point of this detection line, and update the type corresponding to the new end point of this detection line as the ground point cloud, as well as update the length of this detection line; if it is less than or equal to the length of the recorded detection line, no processing is performed. It should be noted that in this way, the outer envelope can be accurately determined based on the ground point cloud data.

[0117] In this embodiment, after determining the outer envelope in the way of the detection line, use the detection line to continue traversing the point cloud of the obstacles in the outer envelope area. Similarly, determine each obstacle point cloud of the obstacles. For each obstacle point cloud, connect each obstacle point cloud with the center point O to determine the included angle between this connection line and the road extension direction, and find the detection line closest to this obstacle point cloud according to this included angle, and determine the distance between this obstacle point cloud and the center point O. If this distance is less than the length of the detection line, use this obstacle point cloud as the end point of this detection line to determine the point cloud closest to the target vehicle, so as to select the point cloud closest to the target vehicle in different second preset directions from each obstacle, and then use these point clouds to determine the drivable area in the second type of partition.

[0118] Please refer to Figure 8 , Figure 8 is a frame schematic diagram of an embodiment of an electronic device provided by the present application. In this embodiment, the electronic device 80 includes a memory 81 and a processor 82 that are coupled to each other.

[0119] The memory 81 stores program instructions, and the processor 82 is configured to execute the program instructions stored in the memory 81 to implement the steps of any of the above method embodiments. In a specific implementation scenario, the electronic device 80 may include, but is not limited to, a microcomputer, a server. In addition, the electronic device 80 may also include mobile devices such as a laptop computer, a tablet computer, etc., which are not limited herein.

[0120] Specifically, the processor 82 is configured to control itself and the memory 81 to implement the steps of any of the above embodiments. The processor 82 may also be referred to as a CPU (Central Processing Unit). The processor 82 may be an integrated circuit chip with signal processing capabilities. The processor 82 may also be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. Additionally, the processor 82 may be implemented jointly by integrated circuit chips.

[0121] Please refer to Figure 9 , Figure 9 which is a schematic framework diagram of the computer-readable storage medium provided by the present application. The computer-readable storage medium 90 of the embodiments of the present application stores program instructions 91, and when the program instructions 91 are executed, the methods provided by any one of the above embodiments and any non-conflicting combinations are implemented. Among them, the program instructions 91 may form a program file and be stored in the above computer-readable storage medium 90 in the form of a software product, so that a computer device (which may be a personal computer, a server, or a network device, etc.) can execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned computer-readable storage medium 90 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 disc, or terminal devices such as a computer, a server, a mobile phone, a tablet, etc.

[0122] In the above solution, after determining the target road boundary from the detection area, the detection area is partitioned based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary. Then, the drivable area in the first type of partition is generated using the first area generation method, and the drivable area in the second type of partition is generated using the second area generation method. Next, the drivable area in the first type of partition and the drivable area in the second type of partition are fused to obtain the drivable area of the target vehicle in the detection area. Since the probability that the first type of partition is within the target road boundary meets the requirements, the generation of the drivable area in the first type of partition is not affected by the target road boundary. Therefore, the use of the first area generation method in this application does not affect the detection effect of the drivable area in the first type of partition. Further, since the processing time of the first area generation method is less than the processing time of the second area generation method, compared with the method of generating the vehicle drivable area using the second area generation method for each partition, this application can not only ensure the detection effect of the drivable area in the entire detection area, but also reduce the detection time to improve the detection efficiency of the drivable area.

[0123] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the methods described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments. For the sake of brevity, it will not be repeated here.

[0124] The descriptions of the above embodiments tend to emphasize the differences between the embodiments. Their similarities or similarities can be referred to each other. For the sake of brevity, they will not be repeated in this article.

[0125] The above are only the implementation manners of this application, and do not limit the patent scope of this application. Any equivalent structure or equivalent process transformation made using the content of the specification and drawings of this application, or directly or indirectly applied to other related technical fields, shall be equally included in the patent protection scope of this application.

Claims

1. A method for detecting a drivable area, characterized in that, The method includes: Determining a target road boundary from several road boundaries in a detection area, where the detection area is within a preset distance range of a target vehicle; Partitioning the detection area based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary, where the probability that the first type of partition corresponding to the target road boundary is within the target road boundary meets the requirements; Generating a drivable area in the first type of partition using a first area generation method, and generating a drivable area in the second type of partition using a second area generation method, where the processing time of the first area generation method is less than that of the second area generation method, and the first area generation method is based on the road boundary passing through the first type of partition to generate the drivable area; Fusing the drivable area in the first type of partition and the drivable area in the second type of partition to obtain the drivable area of the target vehicle in the detection area.

2. The method according to claim 1, characterized in that The partitioning the detection area based on the target road boundary to obtain a first type of partition and a second type of partition corresponding to the target road boundary includes: Finding a target position point in the target road boundary; Taking the line connecting the target position point and the position where the target vehicle is located as the partition line; Using the partition line corresponding to the target road boundary to divide at least part of the detection area into a first type of partition and a second type of partition corresponding to the target road boundary, where the first type of partition corresponding to the target road boundary is on the side of the partition line away from the target road boundary, and the second type of partition is on the side of the partition line close to the target road boundary.

3. The method according to claim 2, wherein The target position point is the farthest position point in the target road boundary and in the preset extension direction of the target road boundary, and the preset extension direction corresponds to the forward direction of the target vehicle; And / or, the using the partition line corresponding to the target road boundary to divide at least part of the detection area into a first type of partition and a second type of partition corresponding to the target road boundary includes: For each target road boundary, taking the area between the reference line of the target vehicle and the partition line corresponding to the target position point of the target road boundary as the first type of partition corresponding to the target road boundary, and determining the area on the side of the partition line close to the target road boundary in the detection area as the second type of partition corresponding to the target road boundary, where the reference line passes through the target vehicle and extends along the length direction of the road where the target vehicle is located.

4. The method according to claim 1, wherein Before determining the target road boundary from several road boundaries in the detection area, it further includes: Obtaining road boundary data of the detection area, where the road boundary data includes boundary position information of the road boundary and the confidence level regarding the boundary position information; The determining the target road boundary from several road boundaries in the detection area includes: Determining the road boundary whose confidence level meets the preset confidence level requirement as the target road boundary.

5. The method according to claim 4, characterized in that The preset confidence requirement is that the confidence is less than or equal to a first confidence threshold and greater than or equal to a second confidence threshold; Before generating the drivable area in the first type of partition by using the first area generation method and generating the drivable area in the second type of partition by using the second area generation method, the method further includes: Obtaining the road boundary with the confidence greater than the first confidence threshold as the first road boundary, and determining the area on the side of the target vehicle close to the first road boundary as the first type of partition; And / or, obtaining the road boundary with the confidence less than the second confidence threshold as the second road boundary, and determining the area on the side of the target vehicle close to the second road boundary as the second type of partition.

6. The method according to claim 5, wherein The determining the area on the side of the target vehicle close to the first road boundary as the first type of partition includes: Determining the area on the first target side of the reference line of the target vehicle in the detection area as the first type of partition, where the first target side is the side close to the first road boundary, and the reference line passes through the target vehicle and extends along the length direction of the road where the target vehicle is located; And / or, the determining the area on the side of the target vehicle close to the second road boundary as the second type of partition includes: Determining the area on the second target side of the reference line of the target vehicle in the detection area as the second type of partition, where the second target side is the side close to the second road boundary.

7. The method according to claim 1, characterized in that, The generating the drivable area in the first type of partition by using the first area generation method includes: Performing target detection on the first type of partition by using a target processing algorithm to obtain a processing result, where the target processing algorithm includes at least one of a target detection algorithm and a target tracking algorithm, and the processing result includes position characterization information of detection frames of several obstacles; Based on the processing result and the road boundary in the first type of partition, obtaining the drivable area in the first type of partition; And / or, the generating the drivable area in the second type of partition by using the second area generation method includes: Performing semantic segmentation on the second type of partition by using a semantic segmentation point algorithm to obtain point cloud data; Based on the point cloud data, obtaining the drivable area in the second type of partition.

8. The method according to claim 7, wherein The position characterization information of the detection frame of the obstacle includes the central position of the center point of the obstacle and size data; the obtaining the drivable area in the first type of partition based on the processing result and the road boundary in the first type of partition includes: For each of the obstacles, determining the position information of each vertex on the detection frame of the obstacle based on the central position and the size data; Based on the magnitude relationship between the angle between each connection line of the detection frame and a first preset direction, determining the vertex pairs of each side belonging to the detection frame, where each connection line is the connection line between each vertex and the center point; Using the position information of the vertex pairs of each side corresponding to each obstacle, selecting the target side closest to the target vehicle in different second preset directions from each of the obstacles; Determine a candidate drivable area of the target vehicle from the first type of partitions based on the target edges corresponding to the respective obstacles; Select, from the candidate drivable area of the target vehicle, an area within the road boundary in the first type of partitions as the drivable area of the first type of partitions.

9. The method according to claim 7, wherein The obtaining the drivable area in the second type of partitions based on the point cloud data includes: Extract ground point clouds belonging to the ground from the point cloud data; Process the ground point clouds to obtain an outer envelope of the ground point clouds; Determine the drivable area in the second type of partitions based on the outer envelope.

10. The method according to claim 1, characterized in that, Before determining the target road boundary from several road boundaries in the detection area, it further includes: Determine a plurality of road boundary points by using the sensing data detected for the detection area; Fit the plurality of road boundary points to obtain the several road boundaries.

11. An electronic device, characterized in that, Comprising a memory and a processor coupled to each other, The memory stores program instructions; The processor is configured to execute the program instructions stored in the memory to implement the method according to any one of claims 1-10.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program instructions capable of being run by a processor, and the program instructions can be executed by the processor to implement the method according to any one of claims 1-10.

Images