Method and device for identifying travelable area of laser radar based on closed scene

Abstract

The invention discloses a laser radar drivable area identification method and device based on a closed scene, and relates to the technical field of intelligent driving, and the method comprises the steps: obtaining corresponding original point clouds around a vehicle body based on a plurality of laser radars installed on a target vehicle, and splicing the original point clouds to obtain corresponding multi-radar splicing point clouds; processing the multi-radar splicing point cloud to obtain corresponding ground point cloud and non-ground point cloud; based on the ground point cloud, the non-ground point cloud and the vehicle body height of the target vehicle, carrying out filtering processing to obtain a vehicle body part point cloud higher than the ground and lower than the vehicle; based on the point cloud of the vehicle body part higher than the ground and lower than the vehicle height and a preset three-dimensional coordinate system, xy-axis plane projection and y-axis plane projection are carried out, and the boundary of the impassable area and the travelable area are obtained through concave packet detection. The point cloud data is obtained based on the preset laser radar around the vehicle body, data processing is carried out, the impassable area boundary and the drivable area are obtained, and an information basis is provided for unmanned driving of a signal interference area.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and specifically to a method and device for identifying a drivable area using a laser radar based on a closed scene. Background Art

[0002] At present, in the field of intelligent driving, especially in closed scenarios such as underground mining areas, GPS positioning signals are often easily interfered with. For example, there is no GPS positioning signal at all in underground mining areas. When there is no GPS positioning signal, intelligent driving vehicles usually use multi-sensor fusion methods such as lidar, IMU and camera for high-precision positioning. However, this positioning method is prone to positioning loss when the sensor is blocked by the vehicle or there are few environmental features. Since there are certain accuracy issues in the visual drivable area, the vehicle needs to plan a forward path based on the laser drivable area.

[0003] In order to meet practical needs, a lidar drivable area recognition technology based on closed scenes is proposed. Summary of the Invention

[0004] In response to the defects in the existing technology, the purpose of this application is to provide a lidar drivable area identification method and device based on closed scenes. The method obtains point cloud data based on the preset lidar around the vehicle body, processes the data, and obtains the boundaries of the inaccessible area and the drivable area, providing information basis for unmanned driving in signal interference areas.

[0005] In order to achieve the above objectives, the technical solution adopted by this application is: In a first aspect, the present application provides a method for identifying a drivable area using a laser radar based on a closed scene, the method comprising the following steps: Based on multiple laser radars installed on the target vehicle, the corresponding original point cloud around the vehicle body is obtained, and then the corresponding multi-radar stitched point cloud is stitched together; Processing the multi-radar stitched point cloud to obtain corresponding ground point cloud and non-ground point cloud; Performing filtering based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; Based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and a preset three-dimensional coordinate system, an xy-axis plane projection is performed, and concave hull detection is performed to obtain the boundary of the impassable area and the drivable area; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

[0006] On the basis of the above technical solution, at least one laser radar is set on the front, back, left and right sides of the body of the target vehicle.

[0007] On the basis of the above technical solution, the processing of the multi-radar stitched point cloud to obtain the corresponding ground point cloud and non-ground point cloud includes the following steps: The multi-radar stitching point cloud is processed based on the patchwork++ method to obtain corresponding ground point cloud and non-ground point cloud.

[0008] On the basis of the above technical solution, the filtering process based on the ground point cloud, the non-ground point cloud and the vehicle height of the target vehicle is performed to obtain the point cloud of the vehicle body portion that is above the ground and below the vehicle height, including the following steps: Performing a two-dimensional grid projection on the ground point cloud according to the xy axis of the three-dimensional coordinate system to obtain the minimum height distance of the corresponding multiple grid ground point clouds; Filtering is performed based on the minimum height distance, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height.

[0009] Based on the above technical solution, after obtaining the boundary of the impassable area and the drivable area, the method further includes the following steps: Based on the boundary of the impassable area and the drivable area, a current local driving trajectory of the vehicle is planned.

[0010] In a second aspect, the present application provides a closed-scene-based laser radar drivable area identification device, the device comprising: A point cloud stitching module is used to obtain the corresponding original point cloud around the vehicle body based on multiple laser radars installed on the target vehicle, and then stitch the corresponding multi-radar stitching point cloud; A point cloud classification module, which is used to process the multi-radar stitching point cloud to obtain corresponding ground point cloud and non-ground point cloud; a point cloud processing module configured to perform filtering processing based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; The area division module is used to perform xy-axis plane projection based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and a preset three-dimensional coordinate system, and obtain the boundary of the impassable area and the drivable area through concave hull detection; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

[0011] On the basis of the above technical solution, at least one laser radar is set on the front, back, left and right sides of the body of the target vehicle.

[0012] On the basis of the above technical solution, the point cloud classification module is also used to process the multi-radar stitching point cloud based on the patchwork++ method to obtain corresponding ground point cloud and non-ground point cloud.

[0013] On the basis of the above technical solution, the point cloud processing module is further used to perform a two-dimensional grid projection on the ground point cloud according to the xy axis of the three-dimensional coordinate system to obtain the minimum height distance of the corresponding multiple grid ground point clouds; The point cloud processing module is further configured to perform filtering based on the minimum height distance, the non-ground point cloud, and the vehicle height of the target vehicle to obtain the point cloud of the vehicle body portion that is above the ground and below the vehicle height.

[0014] On the basis of the above technical solution, the device further includes: A driving planning module is used to plan the current local driving trajectory of the vehicle based on the boundary of the impassable area and the drivable area.

[0015] Compared with the prior art, the advantages of this application are: This application obtains point cloud data based on the preset laser radar around the vehicle body, processes the data, obtains the boundaries of the inaccessible area and the drivable area, and provides information basis for unmanned driving in signal interference areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0017] Figure 1 This is a flowchart of the steps of the closed scene-based lidar drivable area identification method according to an embodiment of the present application; Figure 2 This is a principle flow chart of the closed scene-based lidar drivable area identification method according to an embodiment of the present application; Figure 3 This is a structural block diagram of a closed-scene-based laser radar drivable area identification device according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0019] The embodiments of the present application are further described in detail below with reference to the accompanying drawings.

[0020] The embodiments of the present application provide a method and device for identifying a drivable area using a laser radar based on a closed scene. The method obtains point cloud data based on a laser radar preset around the vehicle body, processes the data, and obtains the boundaries of the inaccessible area and the drivable area, providing information basis for unmanned driving in signal interference areas.

[0021] To achieve the above technical effects, the overall idea of ​​this application is as follows: A method for identifying a drivable area using a laser radar based on a closed scene, the method comprising the following steps: S1. Based on multiple laser radars installed on the target vehicle, obtain the corresponding original point cloud around the vehicle body, and then stitch the corresponding multi-radar stitched point cloud; S2. Process the multi-radar stitched point cloud to obtain the corresponding ground point cloud and non-ground point cloud; S3, performing filtering processing based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; S4, based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and the preset three-dimensional coordinate system, perform xy-axis plane projection, and obtain the boundary of the impassable area and the drivable area through concave hull detection; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

[0022] The embodiments of the present application are further described in detail below with reference to the accompanying drawings.

[0023] First, see Figures 1 and 2 As shown, an embodiment of the present application provides a method for identifying a drivable area using a laser radar based on a closed scene, the method comprising the following steps: S1. Based on multiple laser radars installed on the target vehicle, obtain the corresponding original point cloud around the vehicle body, and then stitch the corresponding multi-radar stitched point cloud; S2. Process the multi-radar stitched point cloud to obtain the corresponding ground point cloud and non-ground point cloud; S3, performing filtering processing based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; S4, based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and the preset three-dimensional coordinate system, perform xy-axis plane projection, and obtain the boundary of the impassable area and the drivable area through concave hull detection; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

[0024] In an embodiment of the present application, point cloud data is obtained based on a preset laser radar around the vehicle body, and data processing is performed to obtain the boundaries of the inaccessible area and the drivable area, providing information basis for unmanned driving in signal interference areas.

[0025] It should be noted that currently, there are two main methods for determining the laser drivable area: a deep learning-based model segmentation method and a conventional segmentation method based on height and laser scanning rules. The deep learning-based model segmentation method directly outputs the drivable area point cloud information based on the input point cloud information after model inference and post-processing of the grid occupation. The conventional segmentation method based on height and laser scanning rules first divides the two-dimensional space into a grid, segments the ground according to a fixed grid height threshold and scanning beam rules, and then generates the drivable area point cloud information based on subsequent processing rules such as expansion.

[0026] Model-based training methods require pre-collection of vehicle point cloud data and segmentation and annotation of single-frame point cloud data. The input single-frame point cloud data and segmentation and annotation results are then fed into the model for training. This method suffers from poor segmentation accuracy if the LiDAR sensor installation location and type are changed without incorporating the changed point cloud data into training. Furthermore, model-based training methods are prone to ground segmentation holes and poor boundary segmentation accuracy. Conventional methods based on height and laser scanning rules consider points with a height Z coordinate below a threshold, a two-dimensional grid bottom elevation difference within the threshold, and no sudden angle changes in the scanning beam direction as potential ground drivable area point clouds. However, when the ground ahead is complex, such as with multiple slopes, uneven terrain, and vertical surfaces, the accuracy of the drivable area segmentation based on the ground rules is poor, resulting in missed segmentations and vehicle stops.

[0027] However, whether based on models or conventional ground segmentation methods, since point clouds may not necessarily scan the entire ground, driving based solely on segmented ground point clouds cannot reflect whether the void area is drivable due to the existence of voids.

[0028] Furthermore, at least one laser radar is provided on each of the front, rear, left and right sides of the target vehicle.

[0029] Furthermore, the processing of the multi-radar stitched point cloud to obtain corresponding ground point cloud and non-ground point cloud includes the following steps: The multi-radar stitching point cloud is processed based on the patchwork++ method to obtain corresponding ground point cloud and non-ground point cloud.

[0030] Furthermore, the filtering process based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height includes the following steps: Performing a two-dimensional grid projection on the ground point cloud according to the xy axis of the three-dimensional coordinate system to obtain the minimum height distance of the corresponding multiple grid ground point clouds; Filtering is performed based on the minimum height distance, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height.

[0031] Furthermore, after obtaining the boundary of the impassable area and the drivable area, the method further includes the following steps: Based on the boundary of the impassable area and the drivable area, a current local driving trajectory of the vehicle is planned.

[0032] Based on the technical solution of the embodiment of the present application, a drivable area method is proposed to reconstruct the boundary of the drivable area based on the laser radar impassable area point cloud. The method mainly uses patchwork++ ground segmentation, post-processing filtering to remove the top and impassable area point cloud projection to reconstruct the vehicle's current drivable area boundary information, so that the downstream planning module can dynamically plan the driving trajectory according to the drivable area information.

[0033] First, the original point cloud data around the vehicle body is obtained based on the lidar sensor installed on the vehicle body. For multiple lidar sensors, the spliced ​​point cloud is obtained based on the calibrated external parameters. The point cloud of the four lidar sensors around the vehicle is spliced ​​together.

[0034] Secondly, based on the ground segmentation method in the paper patchwork++, which mainly includes reflected noise removal (RNR), concentric zone model segmentation (CZM), region vertical plane fitting (R-VPF), region ground fitting (R-GPF), adaptive ground likelihood estimation (A-GLE), and temporal ground recovery (TGR) modules, the ground point cloud of the input spliced ​​point cloud is segmented. The segmented results are ground point cloud and non-ground point cloud.

[0035] Again, the ground point cloud from the previous step may contain holes or be relatively close to the ground. The empty areas cannot reflect the vehicle's drivability. The non-ground point cloud may include other environmental point clouds above the roof. If directly projected, it will indicate that the vehicle is impassable. Therefore, the non-ground point cloud needs to be filtered by vehicle height. The ground point cloud is projected onto a two-dimensional grid with equal resolution (xy) (x is the direction directly in front of the vehicle, y is the direction to the left). The minimum z height distance (z is directly above the vehicle) of each grid ground point cloud is calculated. If the ground point cloud in a grid is empty and the number of ground grids in the surrounding 8 adjacent grids is greater than 2, a weighted average filter is performed based on the minimum height distance of the surrounding grid ground. Then, points with a z height difference greater than the vehicle height in each grid are filtered out, leaving the non-ground point cloud within the vehicle height range.

[0036] Finally, the non-ground point cloud within the vehicle's height range is projected onto a plane using the xy plane to obtain the boundary of the vehicle's impassable area. Since there may be a large number of boundary points, concave hull detection is performed on these points to obtain the impassable area boundary with a lower density. The remaining area is considered to be drivable. The vehicle's current local driving trajectory is subsequently planned based on this boundary.

[0037] The technical solution of the embodiment of the present application performs efficient patchwork++ ground segmentation based on the spliced ​​point cloud of the laser sensor, and reconstructs the vehicle's current drivable area boundary information through post-processing filtering, top removal, and projection of the point cloud of the impassable area, thereby avoiding the distance and empty driving problems caused by only projecting the drivable area, and ensuring the forward movement of the vehicle without positioning information by constructing the boundary of the impassable area.

[0038] Compared with the existing technology, the technical solution of this application has the following technical key features: No data annotation and training are required. The ground and non-ground areas can be segmented by conventional extraction and segmentation of the input point cloud, and the point cloud and boundaries of the inaccessible area can be segmented and extracted. Patchwork++ ground segmentation technology is more effective than simply using a height threshold for segmenting irregular scenes such as slopes and vertical surfaces. Dilation processing easily filters out sharp ground areas but has difficulty processing large ground holes. Projecting impassable areas better restores details in drivable areas than dilation processing.

[0039] Compared with the existing technology, the technical solution of this application has the following technical advantages: No data labeling and training are required, and the requirements for sensor installation location and controller computing power are low. The problem of voids in model segmentation is solved by using information about impassable boundaries. The patchwork++ ground segmentation and post-filtering processing methods are used to improve the segmentation accuracy and adapt to a wide range of scenarios through the inaccessible boundary information.

[0040] As shown in the accompanying drawings Figure 2 As shown, it is the core process of the technical solution of the embodiment of this application. The core process of this application includes: Step 1: Obtain multi-radar stitching point cloud based on the radar installation posture; Step 2: Calculate the ground point cloud and non-ground point cloud processed by the patchwork++ method; Step 3: The ground point cloud calculates the lowest point information of the grid, and after filtering, removes the part of the non-ground point cloud that is higher than the vehicle height according to the vehicle height; Step 4: Remove the non-ground point cloud at vehicle height and perform xy plane projection. After concave hull detection, obtain the sparse boundary of the inaccessible area.

[0041] Compared with the existing technology, the technical solution of this application has the following technical highlights: First, the drivable area detection usually uses the projection of the left ground point cloud to obtain the drivable area range. However, when this method is blocked by an obstacle, there is a large scanning hole on the ground behind it. At this time, the missing ground cannot provide drivability information, and due to the scanning characteristics of the lidar installation, its pure ground scanning is much closer than the obstacle scanning distance. Usually, the effective ground scanning distance is about 50-100m, while the obstacle scanning distance can reach 200m or even farther. Therefore, there are certain limitations in creating a drivable area through pure ground. Therefore, this application further filters the point cloud above the vehicle height through the non-ground point cloud and creates a drivable area through the remaining boundary point cloud.

[0042] Second, vehicle height point filtering uses the previously segmented ground point cloud as a benchmark. The xy plane is divided into equally spaced grids at a resolution of 1.0m. The lowest vertical Z-value is calculated for each ground point within each grid. If a grid has no value and the number of grids with values ​​in the eight adjacent grids is greater than two, a weighted average filter is performed based on the minimum ground height distance of the surrounding grids to calculate the Z-value for the current grid. For grids with values, the non-ground point cloud is projected onto the point cloud of this grid. If the Z-value is greater than the sum of the grid's value and the vehicle height, it is filtered out. If it is less than this, it is retained and considered an obstacle boundary. For grids without values, filtering is performed based on whether the Z-value is higher than the vehicle height in the vehicle sensor coordinate system.

[0043] Third, since the vehicle's obstacle boundary points may project numerous points in the Z direction, sampling of the projected drivable area boundary is necessary to reduce downstream computational burden and data transmission. Direct mesh voxelization would retain a large number of obstacle boundary points outside the inner boundary. Therefore, a concave hull detection (e.g., PCL, Point Cloud Library, pcl::ConcaveHull function) is performed on all projected boundary points around the vehicle to obtain inner boundary points. This reduces unnecessary outer boundary points, thereby reducing the overall boundary point density of the drivable area.

[0044] Second, see Figure 3 As shown, an embodiment of the present application provides a device for identifying a drivable area using a laser radar based on a closed scene, the device comprising: A point cloud stitching module is used to obtain the corresponding original point cloud around the vehicle body based on multiple laser radars installed on the target vehicle, and then stitch the corresponding multi-radar stitching point cloud; A point cloud classification module, which is used to process the multi-radar stitching point cloud to obtain corresponding ground point cloud and non-ground point cloud; a point cloud processing module configured to perform filtering processing based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; The area division module is used to perform xy-axis plane projection based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and a preset three-dimensional coordinate system, and obtain the boundary of the impassable area and the drivable area through concave hull detection; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

[0045] In an embodiment of the present application, point cloud data is obtained based on a preset laser radar around the vehicle body, and data processing is performed to obtain the boundaries of the inaccessible area and the drivable area, providing information basis for unmanned driving in signal interference areas.

[0046] Furthermore, at least one laser radar is provided on each of the front, rear, left and right sides of the target vehicle.

[0047] Furthermore, the point cloud classification module is also used to process the multi-radar stitching point cloud based on the patchwork++ method to obtain corresponding ground point cloud and non-ground point cloud.

[0048] Furthermore, the point cloud processing module is further configured to perform a two-dimensional grid projection on the ground point cloud according to the xy axis of the three-dimensional coordinate system to obtain a minimum height distance of the corresponding plurality of grid ground point clouds; The point cloud processing module is further configured to perform filtering based on the minimum height distance, the non-ground point cloud, and the vehicle height of the target vehicle to obtain the point cloud of the vehicle body portion that is above the ground and below the vehicle height.

[0049] Furthermore, the device further comprises: A driving planning module is used to plan the current local driving trajectory of the vehicle based on the boundary of the impassable area and the drivable area.

[0050] To sum up, the laser radar drivable area identification device based on closed scenes provided in the embodiment of the present application has the same technical principles as the laser radar drivable area identification method based on closed scenes provided in the first aspect in terms of technical problems, technical solutions and technical effects, so they will not be elaborated here.

[0051] In the description of this application, it should be noted that the terms "upper" and "lower" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application. Unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be internal communication between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to the specific circumstances.

[0052] It should be noted that, in this application, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0053] The foregoing is merely a list of specific embodiments of the present application, intended to enable those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the broadest scope consistent with the principles and novel features of the present application.

Claims

1. A method for identifying a drivable area using a laser radar based on a closed scene, characterized in that: The method comprises the following steps: Based on multiple laser radars installed on the target vehicle, the corresponding original point cloud around the vehicle body is obtained, and then the corresponding multi-radar stitched point cloud is stitched together; Processing the multi-radar stitched point cloud to obtain corresponding ground point cloud and non-ground point cloud; Performing filtering based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; Based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and a preset three-dimensional coordinate system, an xy-axis plane projection is performed, and concave hull detection is performed to obtain the boundary of the impassable area and the drivable area; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

2. The method for identifying a drivable area using a laser radar based on a closed scene according to claim 1, wherein: At least one laser radar is arranged on each of the front, rear, left and right sides of the vehicle body of the target vehicle.

3. The closed scene-based laser radar drivable area identification method according to claim 1, characterized in that: The processing of the multi-radar stitched point cloud to obtain corresponding ground point cloud and non-ground point cloud includes the following steps: The multi-radar stitching point cloud is processed based on the patchwork++ method to obtain corresponding ground point cloud and non-ground point cloud.

4. The closed scene-based laser radar drivable area identification method according to claim 1, characterized in that: The filtering process based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of a vehicle body portion that is higher than the ground and lower than the vehicle height includes the following steps: Performing a two-dimensional grid projection on the ground point cloud according to the xy axis of the three-dimensional coordinate system to obtain the minimum height distance of the corresponding multiple grid ground point clouds; Filtering is performed based on the minimum height distance, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height.

5. The closed scene-based laser radar drivable area identification method according to claim 1, characterized in that: After obtaining the impassable area boundary and the drivable area, the method further includes the following steps: Based on the boundary of the impassable area and the drivable area, a current local driving trajectory of the vehicle is planned.

6. A laser radar drivable area recognition device based on a closed scene, characterized in that: The device comprises: A point cloud stitching module is used to obtain the corresponding original point cloud around the vehicle body based on multiple laser radars installed on the target vehicle, and then stitch the corresponding multi-radar stitching point cloud; A point cloud classification module, which is used to process the multi-radar stitching point cloud to obtain corresponding ground point cloud and non-ground point cloud; a point cloud processing module configured to perform filtering processing based on the ground point cloud, the non-ground point cloud, and the vehicle height of the target vehicle to obtain a point cloud of the vehicle body portion that is above the ground and below the vehicle height; The area division module is used to perform xy-axis plane projection based on the point cloud of the vehicle body portion that is above the ground and below the vehicle height and a preset three-dimensional coordinate system, and obtain the boundary of the impassable area and the drivable area through concave hull detection; wherein, The positive direction of the x-axis of the three-dimensional coordinate system is the front direction of the vehicle, the positive direction of the y-axis is the left direction of the vehicle, and the square of the z-axis is the vertical height direction of the vehicle.

7. The closed-scene-based laser radar drivable area recognition device according to claim 6, characterized in that: At least one laser radar is arranged on each of the front, rear, left and right sides of the vehicle body of the target vehicle.

8. The closed-scene-based laser radar drivable area identification device according to claim 6, characterized in that: The point cloud classification module is further used to process the multi-radar stitching point cloud based on the patchwork++ method to obtain corresponding ground point cloud and non-ground point cloud.

9. The closed-scene laser radar drivable area recognition device according to claim 6, characterized in that: The point cloud processing module is further configured to perform a two-dimensional grid projection on the ground point cloud according to the xy axes of the three-dimensional coordinate system to obtain the minimum height distance of the corresponding plurality of grid ground point clouds; The point cloud processing module is further configured to perform filtering processing based on the minimum height distance, the non-ground point cloud, and the vehicle height of the target vehicle to obtain the point cloud of the vehicle body portion that is above the ground and below the vehicle height.

10. The closed scene-based laser radar drivable area recognition device according to claim 6, characterized in that: The device further comprises: A driving planning module is used to plan the current local driving trajectory of the vehicle based on the boundary of the impassable area and the drivable area.

Images