Trailer pose determination method and device, electronic equipment and readable storage medium
By using a dual-filtering method to remove the point clouds of the ground and mop, the point cloud of the main body of the trailer can be accurately reconstructed, solving the problem of inaccurate pose detection of low-profile trailers and achieving higher detection accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-13
- Publication Date
- 2026-03-10
AI Technical Summary
In existing technologies, lidar has difficulty accurately detecting the pose of low-profile trailers, mainly because the trailer body lacks a clear plane, resulting in incomplete point cloud acquisition, which confuses information about other objects in the point cloud and affects detection accuracy.
A dual filtering method is adopted. First, the ground point cloud is filtered out using altitude information, and then the mop point cloud is filtered out using a sliding filter window to obtain the point cloud of the mop body. The pose is determined by fitting the point cloud of the mop body.
It improves the accuracy of trailer pose detection by filtering out interference from ground and mop point clouds, ensuring the purity of point cloud data, accurately reconstructing the main outline of the trailer, and improving detection results.
Smart Images
Figure CN121639788A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present disclosure relates to the technical field of automatic driving, and particularly relates to a method and device for determining a pose of a trailer, an electronic device and a readable storage medium. BACKGROUND
[0002] In the related art, an automatic logistics vehicle includes a tractor and a trailer. In an unmanned automatic logistics vehicle, in order to achieve better path planning and obstacle avoidance operation, the pose of the trailer needs to be detected in real time.
[0003] In a common detection scheme, a laser radar (LiDAR) is used to collect 3D point clouds of the trailer, and the contour of the trailer is detected in the 3D point clouds, and the pose of the trailer is restored in combination with the known size of the trailer and the connection relationship between the trailer and the tractor. Among them, since the laser radar is generally arranged at the tail of the tractor, the laser radar can almost completely collect the point clouds of the front panel of the trailer, and the pose of the front panel is fitted to deduce the pose of the trailer.
[0004] When the trailer is relatively low, the main body of the trailer does not have an obvious plane (such as the front panel), so the laser radar cannot collect complete panel point clouds, and other point clouds will also be mixed into the trailer point clouds, which will result in inaccurate detection results of the pose of the trailer. SUMMARY
[0005] To solve the above technical problems, the present disclosure provides a method and device for determining a pose of a trailer, an electronic device and a readable storage medium, which can improve the accuracy of the detection results of the pose of the trailer.
[0006] In a first aspect, the present disclosure provides a method for determining a pose of a trailer, the method comprising: obtaining an initial point cloud of the trailer; wherein the initial point cloud includes ground point clouds and trailer point clouds, and the trailer point clouds include main body point clouds and mop point clouds; performing a first filtering process on the initial point cloud based on height information in the initial point cloud to filter out the ground point clouds in the initial point cloud and obtain the trailer point clouds; performing a second filtering process on the trailer point clouds based on a sliding filtering window to filter out the mop point clouds in the trailer point clouds and obtain the main body point clouds; determining a plurality of target regions based on a preset reference point, and determining a target point cloud point in the main body point clouds in each target region; wherein the target point cloud point is a point cloud point in the target region that is closest to the preset reference point; determining the pose of the trailer according to the target point cloud point.
[0007] In some embodiments of the present disclosure, optionally, based on the height information in the initial point cloud, the initial point cloud is subjected to a first filtering process to filter out ground point clouds in the initial point cloud to obtain a bucket point cloud, comprising: Based on the height information, the height of each point cloud point in the initial point cloud is obtained; The point cloud points in the initial point cloud with a height lower than a height threshold are marked as ground point clouds; The ground point clouds are filtered out to obtain a bucket point cloud.
[0008] In some embodiments of the present disclosure, optionally, based on a sliding filtering window, the bucket point cloud is subjected to a second filtering process to filter out mop point clouds in the bucket point cloud to obtain a main body point cloud, comprising: Based on a first preset distance, a plurality of sliding filtering windows are generated on the bucket point cloud; wherein each sliding filtering window includes at least part of the point cloud points in the bucket point cloud; The point cloud point with the highest height in each sliding filtering window is marked as a first point cloud point to obtain a first point cloud point set; Each sliding filtering window is moved by a second preset distance in the horizontal direction; wherein the second preset distance is smaller than the first preset distance; The point cloud point with the highest height in each moved sliding filtering window is marked as a second point cloud point to obtain a second point cloud point set; The intersection of the first point cloud point set and the second point cloud point set is determined to filter out the mop point clouds to obtain a main body point cloud.
[0009] In some embodiments of the present disclosure, optionally, based on a first preset distance, a plurality of sliding filtering windows are generated on the bucket point cloud, comprising: A reference axis is generated; wherein the reference axis extends along the vertical direction; A plurality of coaxial annular surfaces passing through the bucket point cloud are generated with the reference axis as the axis, and the distance between any two adjacent annular surfaces in the radial direction of the annular surface is the first preset distance; The area between the two adjacent annular surfaces is determined as a sliding filtering window.
[0010] In some embodiments of the present disclosure, optionally, a plurality of target regions are determined based on a preset reference point, comprising: The main body point cloud is projected onto a preset plane; wherein the preset plane is parallel to the horizontal plane; A preset reference point is generated on the preset plane; wherein the preset reference point is located in front of the bucket; A plurality of first rays are generated on the preset plane with the preset reference point as the endpoint, the first rays extend from the preset reference point to the direction of the main body point cloud, and the included angle between any two adjacent first rays is equal; The region between two adjacent first rays is defined as the target region.
[0011] In some embodiments of this disclosure, optionally, the multiple target regions include multiple first target regions and multiple second target regions; determining the multiple target regions based on preset reference points includes: Project the main point cloud onto a preset plane; wherein the preset plane is parallel to the horizontal plane; A first reference point and a second reference point are generated on a preset plane; wherein both the first reference point and the second reference point are located in front of the trailer. Multiple second rays are generated with the first reference point as the endpoint, and multiple third rays are generated with the second reference point as the endpoint; wherein, the second rays extend from the first reference point toward the main point cloud, and the third rays extend from the second reference point toward the main point cloud, and the included angle between any two adjacent second rays is equal, and the included angle between any two adjacent third rays is equal. The region between two adjacent second rays is defined as the first target region, and the region between two adjacent third rays is defined as the second target region.
[0012] In some embodiments of this disclosure, optionally, determining the pose of the bucket based on the target point cloud includes: The target point cloud is fitted to obtain the target line segment; The position and orientation of the trailer are determined based on the target line segment and the size parameters of the trailer.
[0013] Secondly, embodiments of this disclosure provide a device for determining the position and orientation of a tow bucket, comprising: The acquisition module is used to acquire the initial point cloud of the trailer; the initial point cloud includes the ground point cloud, the main body point cloud, and the mop point cloud; The first filtering module is used to perform a first filtering process on the initial point cloud based on the height information in the initial point cloud, so as to filter out the ground point cloud in the initial point cloud and obtain the bucket point cloud. The second filtering module is used to perform a second filtering process on the bucket point cloud based on a sliding filtering window, so as to filter out the mop point cloud from the bucket point cloud and obtain the main point cloud. The determination module is used to determine multiple target areas based on preset reference points, and to determine a target point cloud point in each target area; wherein the target point cloud point is the point cloud point in the target area that is closest to the preset reference point; and to determine the pose of the trailer based on the target point cloud point.
[0014] Thirdly, embodiments of this disclosure provide an electronic device, including: Memory; Processor; and Computer programs; The computer program is stored in the memory and configured to be executed by the processor to implement the method as provided in the first aspect.
[0015] Fourthly, embodiments of this disclosure provide a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method provided in the first aspect.
[0016] Fifthly, embodiments of this disclosure also provide a computer program product comprising a computer program or instructions that, when executed by a processor, implement the method provided in the first aspect.
[0017] The technical solution provided in this disclosure has the following advantages compared with the prior art: The trailer pose determination method, apparatus, electronic device, and readable storage medium provided in this disclosure employ a dual filtering method. Based on height filtering and sliding filtering, ground point clouds and mop point clouds are filtered out from the initial point cloud of the trailer, resulting in a main point cloud containing only the trailer body. Based on the main point cloud, the target point cloud closest to the tractor is identified. By fitting this target point cloud, an edge of the trailer body can be obtained. Using this edge, combined with information such as the trailer's size and connection method, the main outline of the trailer body can be reconstructed, thereby accurately identifying the trailer's pose information. By filtering out interference from ground and mop point clouds, the accuracy of the trailer pose detection results can be improved. Attached Figure Description
[0018] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.
[0019] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0020] Figure 1 A flowchart of a method for determining the pose of a tractor provided in some embodiments of this disclosure is shown; Figure 2 A side view schematic diagram of the initial point cloud of some embodiments of this disclosure is shown; Figure 3A A schematic diagram of a second filtering process according to some embodiments of this disclosure is shown; Figure 3BA schematic diagram of a second filtering process according to some embodiments of this disclosure is shown; Figure 3C A schematic diagram of a second filtering process according to some embodiments of this disclosure is shown; Figure 4 A schematic diagram of a sliding filter window according to some embodiments of the present disclosure is shown; Figure 5 A schematic diagram of the target area is shown for some embodiments of this disclosure; Figure 6 A schematic diagram of the target area is shown for some embodiments of this disclosure; Figure 7 A schematic diagram of the outline of the trailer is shown for some embodiments of this disclosure; Figure 8 A structural block diagram of a bucket pose determination device according to some embodiments of the present disclosure is shown; Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation
[0021] To better understand the above-mentioned objectives, features, and advantages of this disclosure, the solutions disclosed herein will be further described below. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.
[0022] Numerous specific details are set forth in the following description in order to provide a full understanding of this disclosure, but this disclosure may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only some, and not all, of the embodiments of this disclosure.
[0023] Some of the nouns or terms that appear in the description of the embodiments of this disclosure shall be interpreted as follows: Trailer: A non-powered vehicle used for loading goods. The trailer is connected to the tractor by a tow bar and moves under the traction of the tractor to transport goods.
[0024] Point cloud: A dataset of points in three-dimensional space obtained by discrete sampling of the surface of an object using lidar.
[0025] In some embodiments of this disclosure, a method for determining the pose of a tow bucket is provided. Figure 1 Flowcharts of methods for determining the pose of a tractor provided in some embodiments of this disclosure are shown, such as... Figure 1 As shown, the method includes: S102, Obtain the initial point cloud of the trailer; wherein, the initial point cloud includes the ground point cloud and the trailer point cloud, and the trailer point cloud includes the main body point cloud and the mop point cloud.
[0026] In this embodiment of the disclosure, the pose determination method for the trailer is exemplaryly applied to an automated logistics vehicle, which includes a tractor and a trailer. The trailer is connected to the parking space of the tractor via a tow bar. The parking space of the tractor is also equipped with a lidar, which can collect the initial point cloud of the trailer.
[0027] For lower-profile trailers, when the lidar collects point cloud data, the laser's coverage area includes not only the trailer itself but also the surrounding ground and the mop handle. Here, the unprocessed point cloud data collected by the lidar is defined as the initial point cloud of the trailer.
[0028] S104. Based on the height information in the initial point cloud, the initial point cloud is subjected to a first filtering process to filter out the ground point cloud and obtain the bucket point cloud.
[0029] In this embodiment, the initial point cloud includes a ground point cloud. The ground point cloud refers to the set of point clouds on the ground acquired by the lidar. Since the trailer is moving or parked "on" the ground, the height of the point cloud of the trailer body is higher than the height of the ground point cloud. Therefore, based on the height information in the initial point cloud, a first filtering process is performed on the initial point cloud. Through this first filtering process, the ground point cloud in the initial point cloud can be filtered out, resulting in the trailer point cloud. This trailer point cloud includes both the trailer body point cloud and the mop head point cloud.
[0030] S106, Based on the sliding filter window, a second filtering process is performed on the bucket point cloud to filter out the mop point cloud and obtain the main point cloud.
[0031] In this embodiment of the disclosure, the bucket point cloud includes a main body point cloud and a mop point cloud. When the bucket is relatively low, the height of the mop point cloud may be similar to the height of the bucket point cloud, thus affecting the fitting of the outline of the bucket body.
[0032] To address this issue, this embodiment employs a sliding filter method to perform a second filtering process on the mop bucket point cloud, thereby filtering out the mop point cloud. The remaining point cloud data after filtering out the mop point cloud is the point cloud data of the mop bucket body, also known as the aforementioned body point cloud.
[0033] For example, during the second filtering process, one or more sliding filter windows can be set. By controlling the way the sliding filter windows slide on the mop dot cloud, the main body of the mop bucket and the mop part can be identified, thereby filtering the mop dot cloud.
[0034] S108, determine multiple target regions based on preset reference points, and determine a target point cloud point in the main point cloud within each target region; wherein, the target point cloud point is the point cloud point in the target region that is closest to the preset reference point.
[0035] In this embodiment of the disclosure, after obtaining a main point cloud containing only the main body of the trailer, a preset reference point is set. For example, this preset reference point may be the location of the lidar. Multiple target regions are divided based on the preset reference point, wherein each target region contains at least a portion of the point cloud points from the main point cloud.
[0036] Within each target area, the point cloud point closest to a preset reference point is identified and defined as the target point cloud point. This target point cloud point is the point cloud point in the main point cloud that is closest to the lidar (tractor). By fitting this point cloud point, the edge of the trailer outline closest to the tractor can be obtained.
[0037] S110 determines the position and orientation of the bucket based on the target point cloud.
[0038] In this embodiment of the disclosure, the overall outline of the trailer can be reconstructed by fitting a side of the trailer outline that is close to the tractor vehicle based on the target point cloud points, and combining the known size parameters of the trailer and the connection method between the trailer and the tractor vehicle.
[0039] The trailer pose determination method, apparatus, electronic device, and readable storage medium provided in this disclosure employ a dual filtering method. Based on height filtering and sliding filtering, ground point clouds and mop point clouds are filtered out from the initial point cloud of the trailer, resulting in a main point cloud containing only the trailer body. Based on the main point cloud, the target point cloud closest to the tractor is identified. By fitting this target point cloud, an edge of the trailer body can be obtained. Using this edge, combined with information such as the trailer's size and connection method, the main outline of the trailer body can be reconstructed, thereby accurately identifying the trailer's pose information. By filtering out interference from ground and mop point clouds, the accuracy of the trailer pose detection results can be improved.
[0040] Optionally, in some embodiments of this disclosure, based on the height information in the initial point cloud, a first filtering process is performed on the initial point cloud to filter out ground point clouds from the initial point cloud, obtaining a trailer point cloud, including: Based on the height information, the height of each point in the initial point cloud is obtained; Points in the initial point cloud with heights below a height threshold are marked as ground point clouds; Filter out the ground point cloud to obtain the bucket point cloud.
[0041] In this embodiment of the disclosure, Figure 2 A lateral schematic diagram of the initial point cloud of some embodiments of this disclosure is shown, such as... Figure 2As shown, the initial point cloud 200 includes a trailer point cloud 202 and a ground point cloud 204. Since the trailer is parked or traveling on the ground, the height of the ground point cloud is generally 0.
[0042] Considering factors such as error and ground undulation, a height threshold can be set. During the first filtering process, the set of point cloud points whose height is lower than the height threshold is determined as the ground point cloud, and the ground point cloud is filtered to obtain a bucket point cloud that does not contain the ground point cloud.
[0043] For example, the height threshold can be set according to the height of the trailer chassis. For instance, if the trailer chassis is 30cm off the ground, the height threshold can be 10cm to 15cm.
[0044] This embodiment of the disclosure marks the ground point cloud by comparing the height of each point cloud point in the initial point cloud with a height threshold, thereby achieving accurate filtering of the ground point cloud and improving the purity of the bucket point cloud.
[0045] Optionally, in some embodiments of this disclosure, a second filtering process is performed on the hopper point cloud based on a sliding filtering window to filter out the mop point cloud and obtain the main point cloud, including: Based on a first preset distance, multiple sliding filter windows are generated on the bucket point cloud; wherein each sliding filter window includes at least a portion of the point cloud points in the bucket point cloud; Mark the point cloud point with the highest point cloud height in each sliding filter window as the first point cloud point, and obtain the first set of point cloud points; In the horizontal direction, each sliding filter window is moved a second preset distance; wherein the second preset distance is less than the first preset distance. Mark the point cloud point with the highest point cloud height in each moved sliding filter window as the second point cloud point, and obtain the second point cloud point set. Determine the intersection of the first point cloud set and the second point cloud set to filter out the mop point cloud and obtain the main point cloud.
[0046] In this disclosure embodiment, by way of example, Figure 3A , Figure 3B and Figure 3C A schematic diagram of a second filtering process according to some embodiments of this disclosure is shown, such as... Figure 3A , Figure 3B and Figure 3C As shown, it includes a bucket point cloud 300 and a sliding filter window 302.
[0047] like Figure 3AAs shown, on the bucket point cloud 300, a sliding filter window 302 is defined at intervals of a first preset distance. The point cloud point with the highest height in each sliding filter window 302 is defined as the first point cloud point 304. The set of these first point cloud points 304 is the first point cloud set.
[0048] As shown in 3B, each sliding filter window 302 is moved horizontally, so that each sliding filter window 302 moves a second preset distance in the horizontal direction.
[0049] For example, the second preset distance is less than the first preset distance. For example, the second preset distance is half of the first preset distance.
[0050] For example, the range of the first preset distance is 0.3m to 0.5m.
[0051] For example, when determining the first point cloud set, a sliding filter window can be used to identify the point cloud point with the highest height, and all point cloud points whose height difference with this point cloud point is less than a height difference threshold can be identified as the first point cloud point. For example, the range of the above-mentioned height difference threshold is 0.05m to 0.1m.
[0052] As the sliding filter window moves, the highest point within each sliding filter window may change. For example... Figure 3B As shown, the point cloud point with the highest height is determined again in each sliding filter window 302 and marked as the second point cloud point 306. The set of these second point cloud points 306 is the second point cloud set.
[0053] Since the main body of the trailer is a platform for carrying goods, its height remains constant. However, the mop is typically angled, causing its height to change. Therefore, the intersection of the first and second point cloud sets represents the portion of the point cloud where the point height remains constant; this portion corresponds to the platform of the trailer body. Thus, by taking the intersection of the first and second point cloud sets, we can obtain the main body point cloud after filtering out the mop point cloud, as shown below. Figure 3C As shown, the fitting plane 308 is the upper surface of the main point cloud.
[0054] This embodiment of the present disclosure uses a sliding filter window to filter out the mop portion from the mop point cloud, obtaining the main point cloud of the mop body. This shields the influence of the mop when determining the mop pose, thereby improving the accuracy of mop pose detection.
[0055] Optionally, in some embodiments of this disclosure, multiple sliding filter windows are generated on the bucket point cloud based on a first preset distance, including: Generate a reference axis; wherein the reference axis extends in the vertical direction; With the reference axis as the axis, multiple coaxial annular surfaces are generated that pass through the bucket point cloud, and in the radial direction of the annular surfaces, the distance between any two adjacent annular surfaces is a first preset distance. The region between two adjacent annular surfaces is defined as the sliding filter window.
[0056] In this embodiment of the disclosure, since the bucket point cloud is a three-dimensional point cloud, which includes multiple point cloud points distributed in three-dimensional space, it is necessary to ensure that each sliding filter window corresponds to a spatial region in three-dimensional space when determining the sliding filter window.
[0057] For example, Figure 4 A schematic diagram of a sliding filter window according to some embodiments of this disclosure is shown, such as... Figure 4 As shown, a reference axis 402 perpendicular to the horizontal plane can be generated on one side of the bucket point cloud. Multiple coaxial annular surfaces 404 are generated around the reference axis 402. The radii of these annular surfaces 404 increase in sequence, so that the distance between any two adjacent annular surfaces is equal and is a first preset distance.
[0058] At this point, the control area between two adjacent annular surfaces is a sliding filter window 406.
[0059] This embodiment generates multiple coaxial annular surfaces based on a reference axis, and defines the area between two adjacent annular surfaces as a sliding filter window. The sliding filter window can accurately filter out the mop point cloud, thereby improving the purity of the main point cloud of the mop and improving the accuracy of determining the mop pose.
[0060] In some embodiments of this disclosure, optionally, multiple target regions are determined based on preset reference points, including: Project the main point cloud onto a preset plane; wherein the preset plane is parallel to the horizontal plane; Generate a preset reference point on a preset plane; wherein the preset reference point is located in front of the trailer bucket; Using a preset reference point as the endpoint, multiple first rays are generated on a preset plane. The first rays extend from the preset reference point toward the main point cloud, and the included angle between any two adjacent first rays is equal. The region between two adjacent first rays is defined as the target region.
[0061] In this embodiment of the disclosure, Figure 5 A schematic diagram of the target area of some embodiments of this disclosure is shown, such as Figure 5 As shown, when determining the target area, the main point cloud is first projected onto a preset plane 502, which is a plane parallel to the horizontal plane.
[0062] After projection, a preset reference point 504 is generated on a preset plane. This preset reference point is located in front of the trailer. "In front of the trailer" refers to the direction the trailer faces towards the tractor. For example, the preset reference point 504 can be set within a circular area with a radius of 50cm centered at the connection point between the trailer and the tractor. For example, when the trailer is connected to the tractor, the area between the trailer body and the tractor's cab is defined as the reference area, and the preset reference point is set within this reference area.
[0063] Using a preset reference point 504 as the endpoint, multiple first rays 506 are generated, and the area between two adjacent first rays 506 is defined as the target area 508. The included angle between any two adjacent first rays 506 is equal.
[0064] For example, the included angle between two adjacent first rays ranges from 3° to 30°.
[0065] For example, the angle between two adjacent first rays is positively correlated with the lateral resolution of the lidar.
[0066] The area where the trailer is located is divided into several sectors with specific angles by multiple first rays. Each sector is a target region. Within each target region, the distance from all point cloud points to a preset reference point is calculated, and the point cloud point with the closest distance to the preset reference point within each target region is retained. These point cloud points are the points remaining after filtering out the non-main contour of the trailer, i.e., contour edge points.
[0067] The embodiments disclosed herein can accurately identify the contour edge points of the trailer body, thereby improving the accuracy of trailer pose detection.
[0068] In some embodiments of this disclosure, optionally, the multiple target regions include multiple first target regions and multiple second target regions; determining the multiple target regions based on preset reference points includes: Project the main point cloud onto a preset plane; wherein the preset plane is parallel to the horizontal plane; A first reference point and a second reference point are generated on a preset plane; wherein both the first reference point and the second reference point are located in front of the trailer. Multiple second rays are generated with the first reference point as the endpoint, and multiple third rays are generated with the second reference point as the endpoint; wherein, the second rays extend from the first reference point toward the main point cloud, and the third rays extend from the second reference point toward the main point cloud, and the included angle between any two adjacent second rays is equal, and the included angle between any two adjacent third rays is equal. The region between two adjacent second rays is defined as the first target region, and the region between two adjacent third rays is defined as the second target region.
[0069] Specifically, a first point cloud point is determined in each first target area, and a second point cloud point is determined in each second target area; wherein the first point cloud point is the point cloud point in the first target area that is closest to the second preset reference point, and the second point cloud point is the point cloud point in the second target area that is closest to the third preset reference point.
[0070] In this embodiment of the disclosure, Figure 6 A schematic diagram of the target area of some embodiments of this disclosure is shown, such as Figure 6 As shown, when determining the target area, the main point cloud is first projected onto a preset plane 602, which is a plane parallel to the horizontal plane.
[0071] After projection is completed, two preset reference points are generated on the preset plane: a first preset reference point 604A and a second preset reference point 604B. Both preset reference points are located in front of the trailer bucket. "In front of the trailer bucket" refers to the direction in which the trailer bucket faces the tractor unit.
[0072] For example, both the first preset reference point and the second preset reference point can be set within a circular area with a radius of 50cm centered on the connection point between the mop and the tractor, and the minimum distance between the first preset reference point and the main point cloud is equal to the minimum distance between the second preset reference point and the main point cloud.
[0073] For example, when the trailer is connected to the tractor, the area between the trailer body and the front of the tractor is defined as the reference area. Then, the first preset reference point and the second preset reference point are both set in the reference area, and the first preset reference point and the second preset reference point are respectively located on the left and right sides of the mop.
[0074] The target area includes the first target area 608A and the second target area 608B.
[0075] Using the first preset reference point 604A as the endpoint, multiple second rays 606A are generated, and the area between two adjacent second rays 606A is defined as the first target area 608A. The included angle between any two adjacent second rays 606A is equal.
[0076] Using the second preset reference point 604B as the endpoint, multiple third rays 606B are generated, and the region between two adjacent third rays 606B is defined as the second target region 608B. The included angle between any two adjacent third rays 606B is equal.
[0077] For example, the angle between two adjacent second rays ranges from 3° to 30°. The angle between two adjacent third rays also ranges from 3° to 30°.
[0078] For example, the angle between two adjacent second rays is positively correlated with the lateral resolution of the lidar. The angle between two adjacent third rays is also positively correlated with the lateral resolution of the lidar.
[0079] The area where the trailer is located is divided into several sectors with specific angles by multiple second rays. Each sector is a first target region. Within each first target region, the distances of all point cloud points to a preset reference point are calculated, and the first point cloud point within each target region that is closest to the preset reference point is retained.
[0080] The area where the trailer is located is further divided into sectors with specific angles using multiple third rays. Each sector divided by the third rays is a second target region. Within each second target region, the distances from all point cloud points to a preset reference point are calculated, and the second point cloud point within each target region that is closest to the preset reference point is retained.
[0081] The combination of the first and second point cloud points is determined as the final target point cloud point.
[0082] The embodiments disclosed herein use multi-angle reference points for radial filtering, which can effectively filter out non-main body contour points while enhancing the ability to capture point cloud points on the edge of the trailer, thereby improving the adaptability of the pose determination method.
[0083] In some embodiments of this disclosure, optionally, determining the pose of the bucket based on the target point cloud includes: The target point cloud is fitted to obtain the target line segment; The position and orientation of the trailer are determined based on the target line segment and the size parameters of the trailer.
[0084] In this embodiment of the disclosure, the target point cloud point is the point cloud point within each target area that is closest to a preset reference point, i.e., the edge point of the contour. After fitting these target point cloud points, the resulting target line segment is the edge of the trailer contour that is closest to the tractor.
[0085] For example, Figure 7 A schematic diagram of the bucket outline of some embodiments of this disclosure is shown, such as Figure 7 As shown, target line segment 704 is obtained by fitting target point cloud points 702. By combining target line segment 704 with known dimensions of the trailer and the connection method between the trailer and the tractor, the overall outline of the trailer can be reconstructed, thereby obtaining the pose information of the trailer.
[0086] The embodiments disclosed herein can accurately identify the outline segments of the trailer, thereby accurately obtaining the position and orientation of the trailer.
[0087] In some embodiments of this disclosure, a device for determining the position and orientation of a tractor is also provided. Figure 8 Structural block diagrams of a bucket pose determination device according to some embodiments of this disclosure are shown, such as... Figure 8 As shown, the bucket position determination device 800 includes: The acquisition module 802 is used to acquire the initial point cloud of the trailer; wherein, the initial point cloud includes the ground point cloud, the main body point cloud and the mop point cloud; The first filtering module 804 is used to perform a first filtering process on the initial point cloud based on the height information in the initial point cloud, so as to filter out the ground point cloud in the initial point cloud and obtain the bucket point cloud. The second filtering module 806 is used to perform a second filtering process on the bucket point cloud based on a sliding filtering window, so as to filter out the mop point cloud from the bucket point cloud and obtain the main point cloud. The determination module 808 is used to determine multiple target areas based on a preset reference point, and to determine a target point cloud point in each target area; wherein the target point cloud point is the point cloud point in the target area that is closest to the preset reference point; and to determine the pose of the trailer based on the target point cloud point.
[0088] The trailer pose determination method, apparatus, electronic device, and readable storage medium provided in this disclosure employ a dual filtering method. Based on height filtering and sliding filtering, ground point clouds and mop point clouds are filtered out from the initial point cloud of the trailer, resulting in a main point cloud containing only the trailer body. Based on the main point cloud, the target point cloud closest to the tractor is identified. By fitting this target point cloud, an edge of the trailer body can be obtained. Using this edge, combined with information such as the trailer's size and connection method, the main outline of the trailer body can be reconstructed, thereby accurately identifying the trailer's pose information. By filtering out interference from ground and mop point clouds, the accuracy of the trailer pose detection results can be improved.
[0089] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. The electronic device provided in an embodiment of this disclosure can execute the processing flow provided in the request processing method embodiment, such as... Figure 9 As shown, the electronic device 90 includes: a memory 91, a processor 92, a computer program, and a communication interface 93; wherein the computer program is stored in the memory 91 and is configured to be executed by the processor 92 as described above for determining the pose of the tow bucket.
[0090] In addition, this disclosure also provides a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the pose determination method for the tow bucket described in the above embodiments.
[0091] Furthermore, this disclosure also provides a computer program product, which includes a computer program or instructions that, when executed by a processor, implement the pose determination method for the tow bucket as described above.
[0092] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0093] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. 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 this disclosure. Therefore, this disclosure is not to be limited to the embodiments described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method of pose determination of a drag reducing device, characterized in that, The method comprises: obtaining an initial point cloud of the mop; wherein the initial point cloud comprises a ground point cloud and a mop point cloud, and the mop point cloud comprises a main body point cloud and a mop handle point cloud; performing first filtering processing on the initial point cloud based on height information in the initial point cloud to filter out the ground point cloud in the initial point cloud and obtain a mop point cloud; performing second filtering processing on the mop point cloud based on a sliding filtering window to filter out the mop handle point cloud in the mop point cloud and obtain a main body point cloud; determining a plurality of target regions based on a preset reference point, and determining a target point cloud point in the main body point cloud in each target region; wherein the target point cloud point is a point cloud point closest to the preset reference point in the target region; determining a pose of the mop according to the target point cloud point.
2. The method of claim 1, wherein, The method comprises: obtaining an initial point cloud of the mop; wherein the initial point cloud comprises a ground point cloud and a mop point cloud, and the mop point cloud comprises a main body point cloud and a mop handle point cloud; performing first filtering processing on the initial point cloud based on height information in the initial point cloud to filter out the ground point cloud in the initial point cloud and obtain a mop point cloud; performing second filtering processing on the mop point cloud based on a sliding filtering window to filter out the mop handle point cloud in the mop point cloud and obtain a main body point cloud.
3. The method of claim 1, wherein, The method comprises: generating a plurality of sliding filtering windows on the mop point cloud based on a first preset distance; wherein each of the sliding filtering windows comprises at least part of the point cloud points in the mop point cloud; marking a point cloud point with the highest height in each of the sliding filtering windows as a first point cloud point to obtain a first point cloud point set; moving each of the sliding filtering windows by a second preset distance in a horizontal direction; wherein the second preset distance is smaller than the first preset distance; marking a point cloud point with the highest height in each of the moved sliding filtering windows as a second point cloud point to obtain a second point cloud point set; determining an intersection of the first point cloud point set and the second point cloud point set to filter out the mop handle point cloud and obtain the main body point cloud.
4. The method of claim 3, wherein, The method comprises: generating a reference axis; wherein the reference axis extends in a vertical direction; generating a plurality of coaxial annular surfaces passing through the mop point cloud with the reference axis as an axis, and the distance between any two adjacent annular surfaces in a radial direction of the annular surface is the first preset distance; determining a region between the two adjacent annular surfaces as the sliding filtering window.
5. The method according to any one of claims 1 to 4, characterized in that, The method comprises: projecting the main body point cloud onto a preset plane; wherein the preset plane is parallel to a horizontal plane; generating the preset reference point on the preset plane; wherein the preset reference point is located in front of the mop. A plurality of first rays are generated on the preset plane with the preset reference point as an end point, the first rays extend from the preset reference point to the direction of the main body point cloud, and the included angle between any two adjacent first rays is equal; The area between the two adjacent first rays is determined as the target area.
6. The method according to any one of claims 1 to 4, characterized in that, The plurality of target areas include a plurality of first target areas and a plurality of second target areas; the plurality of target areas are determined based on a preset reference point, including: The main body point cloud is projected onto a preset plane; wherein the preset plane is parallel to the horizontal plane; A first reference point and a second reference point are respectively generated on the preset plane; wherein the first reference point and the second reference point are located in front of the drag bucket; A plurality of second rays are generated with the first reference point as an end point, and a plurality of third rays are generated with the second reference point as an end point; wherein the second rays extend from the first reference point to the direction of the main body point cloud, the third rays extend from the second reference point to the direction of the main body point cloud, the included angle between any two adjacent second rays is equal, and the included angle between any two adjacent third rays is equal; The area between the two adjacent second rays is determined as the first target area, and the area between the two adjacent third rays is determined as the second target area.
7. The method according to any one of claims 1 to 4, characterized in that, The pose of the drag bucket is determined according to the target point cloud point, including: The target point cloud point is fitted to obtain a target line segment; The pose of the drag bucket is determined according to the target line segment and the size parameter of the drag bucket.
8. A device for determining the pose of a dragline, characterized in that Including: An acquisition module is configured to acquire an initial point cloud of a drag bucket; wherein the initial point cloud includes a ground point cloud, a main body point cloud, and a mop point cloud; A first filtering module is configured to perform first filtering processing on the initial point cloud based on height information in the initial point cloud, to filter out the ground point cloud in the initial point cloud, and obtain a drag bucket point cloud; A second filtering module is configured to perform second filtering processing on the drag bucket point cloud based on a sliding filtering window, to filter out the mop point cloud in the drag bucket point cloud, and obtain a main body point cloud; A determination module is configured to determine a plurality of target areas based on a preset reference point, and determine a target point cloud point in each target area; wherein the target point cloud point is the point cloud point in the target area closest to the preset reference point; and determine the pose of the drag bucket according to the target point cloud point.
9. An electronic device, comprising: Including: a memory; a processor; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method of any one of claims 1-7.
10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1-7.