A method for truck bed contour sizing and load position correction
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- YANGTZE RIVER DELTA HART ROBOT IND TECH RES INST
- Filing Date
- 2022-12-27
- Publication Date
- 2026-06-26
Smart Images
Figure CN116184354B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of measurement and positioning in warehousing and logistics, and particularly to a method for measuring the outline dimensions of a truck body and correcting the loading and unloading position based on lidar. Background Technology
[0002] In the future unmanned warehousing and logistics process, one of the important links is that robots automatically load and unload goods in trucks according to the loading and unloading tasks. Therefore, the outline dimensions of the truck body and the local outline deformation of the truck body will play a crucial role in the safety of robots loading and unloading goods in the truck body.
[0003] Currently, some devices on the market obtain vehicle size information through laser scanning, but these devices occupy a large space and have significant errors. Other devices obtain vehicle cargo box size information through cameras, but their structures are overly complex and require large spaces. For longer vehicles, they often have limitations and cannot measure dimensions. Furthermore, these devices often have a limited number of cameras, resulting in errors in the measured vehicle cargo box dimensions and position information, which can lead to unexpected situations during bagging and loading. Moreover, current measurement technologies are primarily designed for measuring the 3D box-like contours of trucks and other similar vehicles, mainly for automating the loading of bagged and boxed materials. Technology for measuring the pure 2D contour dimensions of truck cargo boxes remains very rare.
[0004] For example, in the system and method for AGV to store and retrieve goods from a box truck, published in CN114265374A, the solution uses the straight line of the box, the distance between the left and right sides of the box, and the four corner points of the laser point cloud as the pose data of the key points of the box. Moreover, the actual length L and width W of the box are read from the truck type database. However, in the actual loading and unloading tasks, the length and width of the box read from the database can only be used as a reference. The wrinkles of the inner wall of the box and the local deformation of the box will cause the actual box to be not the theoretical rectangle. This solution uses the length L and width W of the box read from the truck type database to deploy the loading and unloading tasks. However, the inner wall of the truck box is wrinkled and uneven, and there are local steel plate deformations inside the box. Especially for trucks with long boxes, the internal situation of the box is complicated. It is not possible to deploy the loading and unloading tasks of the box solely based on the recorded length and width data.
[0005] 2) During actual loading and unloading operations, due to deformation of the interior of the carriage, it cannot be guaranteed that the pallets and other goods will not collide or interfere with the wrinkled interior walls of the carriage, thus causing safety issues. Summary of the Invention
[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for measuring the outline dimensions of a truck body and correcting the loading and unloading position. The method measures the outline dimensions of the truck body using a lidar system, rather than using truck dimensions from a database as in the prior art. The measurement results are more accurate and can provide basic parameters for automatic loading and unloading of goods. Furthermore, it can achieve the correction of the loading and unloading position.
[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a method for measuring the outline dimensions of a truck cargo box and correcting the loading and unloading position, characterized by comprising the following steps:
[0008] (1) Select the location for measuring the truck body profile;
[0009] (2) The outline of the truck body is repeatedly scanned by lidar to collect multiple frames of effective point cloud;
[0010] (3) Fusion of point cloud contour data of truck body;
[0011] (4) Straight-line fitting of the length and width of the carriage;
[0012] (5) Obtain the outline line corresponding to the interior space of the carriage, and calculate the actual length and width of the carriage interior accordingly.
[0013] Step (6) During the loading and unloading task of the robot, the area around the target point of the robot loading and unloading is scanned in real time to determine whether there is a collision, and the target point is corrected based on the collision.
[0014] Step (1) includes: parking the robot in the truck compartment contour measurement area according to the truck external robot navigation and positioning system, and ensuring that the scanning field of the LiDAR at the truck compartment contour measurement position where the robot is parked covers the interior area of the truck compartment.
[0015] Step (2) includes: scanning the outline of the truck body multiple times with at least one lidar to obtain n frames of truck body outline point cloud; performing point cloud preprocessing on each frame of point cloud data obtained from the multiple lidar point cloud data obtained by scanning, and extracting the effective truck body interior point cloud data of each frame of point cloud data.
[0016] Point cloud preprocessing is performed on each frame of the multi-frame LiDAR point cloud data obtained by scanning, including point cloud filtering, region growing, and contour interpolation.
[0017] Point cloud filtering: Perform point cloud filtering on each of the n frames of the truck bed outline point cloud, calculate the distance l from each point p(x, y) in each frame of the point cloud to the origin of the point cloud coordinate system of that frame, and determine the maximum possible effective distance l for actual truck bed parking. max and the maximum effective value of radar ranging l maxrangeFilter out all unrealistic and unreasonable noise points in the point cloud;
[0018] Region Growth: Based on a fixed constraint angle θ of the radar scan of the interior contour of the carriage, a small part of the carriage interior point cloud c is directly extracted from the filtered point cloud. Then, by calculating the Euclidean distance between adjacent points, the region growth is performed from the small part of point cloud c to both sides to complete all the contour point clouds inside the carriage.
[0019] Contour interpolation: Interpolate the complete contour point cloud c inside the carriage. Calculate the Euclidean distance d between two points in point cloud c according to the order of points in the laser frame scan, and select the maximum constraint distance d between points. th We obtain the minimum number of supplementary points k between two points, and then uniformly fill k points according to the minimum number of supplementary points to obtain all valid point cloud data for each frame.
[0020] Step (3) includes:
[0021] The point cloud data extracted from multiple LiDAR frames and the corresponding robot poses for each frame are fused. Each LiDAR scan corresponds to the robot pose for that frame. n mapping scans will acquire n pairs of LiDAR scans and robot poses. Based on the robot poses, the point clouds from the n LiDAR frames are unified to the same reference coordinate system O. ref Below, all points p in the point cloud of the i-th frame i Unify to coordinate system O ref The transformed point cloud p is obtained below. i_trans :
[0022] p i_trans =T(pose) i -pose o )*p i (i = 1, ..., n)
[0023] For n sets of point clouds p unified to a reference coordinate system i_trans All points in the laser frame scanning sequence are fused according to the index, and all points in the same sequence are fused to form a single-frame fused point cloud p. map :
[0024]
[0025] In step (4), the point cloud p is fused from a single frame. mapThe minimum 2D bounding box (OBB) of the truck body contour map point cloud is calculated. Based on the major and minor axis poses of the OBB bounding box, the approximate pose of the truck body is determined. The major and minor axis poses are used as pose constraints for line fitting. The Random Sample Consensus (RANSAC) method is used to continuously perform parallel line fitting and Euclidean clustering on the truck body contour map point cloud to obtain the parallel line poses on both sides of the truck body and the vertical line pose of the truck body head.
[0026] Step (5) includes:
[0027] 1) After obtaining the pose parameters of the parallel lines on both sides of the truck bed and the pose parameters of the vertical line at the head of the truck bed, the three lines are moved parallel to the inside of the truck bed to obtain three new lines, line′, until the single-frame fused point cloud p map All contour points corresponding to the three straight lines are located outside the carriage contour described by the new three straight lines line′. The size of the carriage corresponding to the three straight lines line′ is used as the effective size of the actual interior space of the carriage.
[0028] 2) Define the three new outline lines of the truck body: line′ and parallel lines on both sides. 1、2 The intersection points of line′3, which is perpendicular to the head of the carriage, are pt1 and pt2, respectively. Using line′ 1、2 Line '3 divides the outer contour area of the carriage into three regions: region 1, region 2, and region 3, and moves the parallel straight lines on both sides away from line '3'. 1、2 The resulting regions are region1 and region2. The car body outline points corresponding to region1 and region2 are projected onto line′1 and line′2 respectively. The points pt3 and pt4, which are farthest from pt1 and pt2, are the other two points pt3 and pt4 at the outermost edge of the rear of the car.
[0029] 3) Calculate the actual length L and width D inside the carriage based on the four points pt1, pt2, pt3, and pt4:
[0030]
[0031]
[0032] In step (6), during the loading and unloading task of the robot, a collision avoidance control area is set up in front of the robot. As the robot slowly approaches the placement point, the presence of the vehicle outline point cloud in the collision avoidance area is monitored in real time. If the vehicle outline point cloud is found in the collision avoidance control area in front of the robot, it is determined that there is a collision between the collision avoidance control area in front of the robot and the vehicle outline. At this time, it is determined whether the actual point cloud outline around the cargo deployment point will collide with the cargo. If there is a collision, the loading and unloading position is adjusted by moving inward along the direction of line′3, which is perpendicular to the head of the vehicle.
[0033] The advantages of this invention are as follows: by using parallel straight lines to extract the straight lines between the sides and the head of the truck body, and correcting the actual usable loading and unloading area inside the truck based on the point cloud of the sides and the head area, more accurate internal length and width dimensions of the truck body are provided, which helps to rationally deploy the goods before loading and unloading, reduces the difficult adjustment of the robot pose in a small area during the later real-time navigation and positioning process, improves the smoothness of the actual deployment task process, and reduces the difficulty of later real-time navigation and positioning.
[0034] By setting a real-time anti-collision control zone, the actual point cloud contour around the cargo placement location is detected in real time. Once the possibility of collision between the cargo placement area and the local contour of the carriage is detected, the cargo is moved inward along the direction of line′3, which is perpendicular to the head of the carriage, until the cargo can be safely placed. This ensures that the pallets and other cargo will not collide with the folded inner wall of the carriage, greatly reducing the risk of cargo collision and reducing the manufacturer's logistics loss costs. Attached Figure Description
[0035] The following is a brief explanation of the contents of each of the accompanying drawings and the markings in the drawings:
[0036] Figure 1 This is a flowchart of the method for dimensional measurement and cargo loading / unloading position correction of the present invention;
[0037] Figure 2 This is a schematic diagram illustrating the principle of region growing preprocessing for point clouds in this invention;
[0038] Figure 3 This is a schematic diagram of the point cloud coordinate system;
[0039] Figure 4 This is a schematic diagram illustrating the principle of the present invention: moving three straight lines parallel to the inside of a truck bed to obtain three new straight lines.
[0040] Figure 5 This is a schematic diagram illustrating the principle of projecting the outline points of the rear of the carriage according to the present invention.
[0041] Figure 6 This is a schematic diagram illustrating the collision correction principle during loading and unloading of goods according to the present invention. Detailed Implementation
[0042] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and the description of the preferred embodiments.
[0043] This technical solution proposes a method for measuring the outline dimensions of a truck body and correcting the loading and unloading position based on lidar:
[0044] First, park the robot in a fixed position outside the truck to ensure that the LiDAR can scan the entire outline of the truck's interior.
[0045] Secondly, by scanning the outline of the truck body multiple times with at least one lidar, point cloud preprocessing such as point cloud filtering, region growing, and outline interpolation is performed on each frame of point cloud data obtained from the scanning to extract the effective internal point cloud data of the truck body in each frame.
[0046] Next, the point cloud data is fused together with all the valid point cloud data extracted from the multi-frame LiDAR and the robot pose corresponding to each frame of valid data.
[0047] Then, parallel line fitting and Euclidean clustering are performed on the fused point cloud of the truck compartment to obtain parallel lines on both sides of the truck compartment and vertical lines at the head of the truck compartment. Based on the point cloud thickness on both sides of each line, high-precision measurement of the length and width of the truck compartment is achieved.
[0048] Finally, during the robot's loading and unloading tasks, the area around the target point of the robot's loading and unloading is scanned in real time to correct the target point and prevent safety accidents such as collisions during the loading and unloading process.
[0049] The flowchart of the method in this application is as follows: Figure 1 As shown, the specific implementation plan is as follows:
[0050] (1) Select the location for measuring the truck body profile
[0051] The robot is positioned within the truck bed contour measurement area using the external robot navigation and positioning system, ensuring that the LiDAR scanning field of view at the truck bed contour measurement location where the robot is positioned covers the interior area of the truck bed.
[0052] (2) Collect multiple effective point clouds by repeatedly scanning the truck body outline using a lidar: 1) Scan the truck body outline multiple times using at least one lidar to obtain n frames of truck body outline point clouds; 2) Perform point cloud filtering on each of the n frames of truck body outline point clouds, calculate the distance l from each point p(x, y) in each frame of point cloud to the origin of the coordinate system of that frame of point cloud, and determine the maximum possible effective distance l for actual truck body parking. max and the maximum effective value of radar ranging l maxrange Filter out all unrealistic and unreasonable noise; filtering refers to filtering out noise greater than the maximum effective distance. max and the maximum effective value of radar ranging l maxrange The point; filtering is mainly for points exceeding l max Data points of value; l maxrange The maximum effective value is an inherent property of the radar.
[0053]
[0054] l <l max
[0055] l <l maxrange
[0056] 3) Based on a fixed constraint angle θ of the radar scanned interior contour of the carriage, a small portion of the carriage interior point cloud c is directly extracted from the filtered point cloud. Then, by calculating the Euclidean distance between adjacent points, the small portion of point cloud c is grown outwards to both sides to complete the entire interior contour point cloud of the carriage. Figure 2 Indication.
[0057] 4) Interpolate the complete outline point cloud c inside the carriage, and perform point cloud c according to the points (p1, p2, p3, ..., p) in the laser frame scan. n Calculate the Euclidean distance d between two points in the following order, and select the maximum constraint distance d between the points. th Find the minimum number of points k between two points, and fill k points (p1, p2, ..., p) evenly according to the minimum number of points k. k This process obtains all valid point cloud data for each frame, thereby ensuring high consistency in the density of the contour point cloud region and improving the accuracy and precision of subsequent point cloud straight line fitting.
[0058] d i，i+1 =||p i -p i+1 ||(i=1,…,n-1)
[0059]
[0060]
[0061] Where, p id represents the coordinates of the i-th point in the laser frame scan. i，i+ 1 represents the Euclidean distance between the i-th point and the (i+1)-th point, and k represents the distance between them. i，i+1 Let d represent the minimum number of points to fill between the i-th point and the (i+1)-th point. th This represents the maximum constraint distance between points.
[0062] (3) Fusion of truck body point cloud contour data
[0063] The point cloud data extracted from multiple LiDAR frames and the corresponding robot pose for each frame are fused. Each LiDAR scan corresponds to the robot pose for that frame. n mapping scans will acquire n pairs of LiDAR scans and robot poses. Based on the robot pose, the point clouds from the n LiDAR frames are unified to the same reference coordinate system O. ref (The robot's pose during the first scan can be used as the reference coordinate system O) ref ), where p is a point cloud in the i-th frame. i Unify to coordinate system O ref The transformed point cloud p is obtained below. i_trans
[0064] p i_trans =T(pose) i -pose o )*p i (i = 1, ..., n)
[0065] Wherein, T(pose) i -pose o The robot pose relative to the reference coordinate system O during each mapping operation is... ref homogeneous transformation matrix
[0066] For n sets of point clouds p unified to a reference coordinate system i_trans All points in the laser frame scanning sequence are fused according to the index, and all points in the same sequence are fused to form a single-frame fused point cloud p. map .
[0067]
[0068] Where, p map，index It refers to the coordinates of the point with index 'index' in the fused single-frame point cloud, p i_trans，index The coordinates of the point with index 'index' in the single-frame point cloud before fusion.
[0069] (4) Straight-line fitting of the length and width of the carriage
[0070] Based on the single-frame fused point cloud p mapThe minimum 2D bounding box (OBB) of the truck body contour map point cloud is calculated. Based on the major and minor axis poses of the OBB bounding box, the approximate pose of the truck body can be determined. The major and minor axis poses are used as pose constraints for straight line fitting. The Random Sample Consensus (RANSAC) method is used to continuously perform parallel straight line fitting and Euclidean clustering on the truck body contour map point cloud to obtain the parallel straight line poses on both sides of the truck body and the vertical straight line pose of the truck body head.
[0071] (5) Measure and calculate the actual length and width dimensions of the carriage.
[0072] 1) Given the pose parameters of the parallel lines on both sides of the truck bed and the pose parameters of the vertical line at the front of the truck bed, move the three lines parallel to the inside of the truck bed to obtain three new lines line′, until the single-frame fused point cloud p map All contour points corresponding to the three straight lines are located outside the carriage contour described by the new three straight lines line′. At this point, the three straight lines correspond to the effective size of the actual interior space of the carriage. Figure 3 Indication.
[0073] 2) Define the three new outline lines of the truck body: line′ and parallel lines on both sides. 1、2 The intersection points of line′3, which is perpendicular to the head of the carriage, are pt1 and pt2, respectively. Using line′ 1、2 And line′3 will divide the external outline area of the carriage according to Figure 5 After being divided, it becomes new region1, region2 and region s By projecting the car body contour points corresponding to region1 and region2 onto line′1 and line′2 respectively, we can obtain two additional points pt3 and pt4 at the outermost edge of the rear of the car, as shown below. Figure 5 Indication.
[0074] 3) Calculate the actual length L and width D inside the carriage:
[0075]
[0076]
[0077] (6) Correction of loading and unloading positions
[0078] Before the robot enters the truck bed to perform loading and unloading tasks, the placement of items such as palletized goods has been rationally planned and flexible path information has been established based on the truck bed's dimensions and contour data. The robot scheduling system controls the robot to pick up the goods according to the flexible path information and move them towards the placement point. During this process, the closer the goods are to the placement point, the higher the probability of collision with protrusions caused by unevenness or deformation of the truck bed's side walls. Therefore, a collision avoidance control zone is drawn in front of the robot. As the robot slowly approaches the placement point, the presence of the truck bed's contour point cloud within the collision avoidance zone is monitored in real time. If a collision is detected between the collision avoidance control zone in front of the robot and the truck bed's contour, it should be promptly determined whether the actual point cloud contour around the placement point will collide with the goods. If a collision occurs, the loading and unloading position should be adjusted inwards, moving the goods inwards along a straight line (line′3) perpendicular to the truck bed's head until the goods can be safely placed. Figure 6 Indication.
[0079] The key points and protections of this application are as follows: This invention re-divides the actual internal area of the truck bed using the truck bed contour point cloud and fitted straight lines, and calculates the actual length and width dimensions of the truck bed based on the actual internal area. Furthermore, based on the reasonable placement of palletized goods during loading and unloading tasks, this invention detects the actual point cloud contour around the goods placement point in real time before placement, sets a collision avoidance control zone, and adjusts the goods placement point in a timely manner when collisions are likely to occur due to truck bed deformation or other factors around the goods placement point to prevent collision accidents and reduce safety hazards.
[0080] The solution proposed in this application has the following advantages:
[0081] 1) Due to the wrinkles and unevenness of the truck bed's inner walls, and the presence of localized steel plate deformation, especially for trucks with long beds, the internal conditions are complex. Therefore, simply relying on the length and width data of the truck bed for loading and unloading tasks is insufficient. This solution extracts the straight lines between the sides and the head of the truck bed using parallel straight lines and corrects the usable loading and unloading area inside the truck based on the point cloud data of the sides and head region. It also provides more accurate dimensions of the truck bed's internal length and width, facilitating the rational deployment of goods before loading and unloading, reducing the difficulty of adjusting the robot's pose in small areas during later real-time navigation and positioning, improving the smoothness of the actual deployment process, and reducing the challenges of later real-time navigation and positioning.
[0082] 2) Due to deformation within the truck bed, the condition of the truck's interior walls around the current target point should be monitored in real time to prevent pallets and other goods from colliding with the wrinkled interior walls and causing safety issues. As the robot approaches the cargo placement point, the laser scanning point cloud becomes denser, providing a clearer perception of the truck bed's outline in local areas. This solution establishes a real-time collision avoidance control zone to monitor the more detailed and clear actual point cloud outline around the cargo placement location. If a potential collision is detected between the cargo placement area and the truck bed's local outline, the cargo is moved inward along a straight line (line′3) perpendicular to the truck bed's front until it can be safely placed. This ensures that pallets and other goods will not collide with the wrinkled interior walls, significantly reducing the risk of collision and lowering the manufacturer's logistics losses.
[0083] Obviously, the specific implementation of this invention is not limited to the above-described manner. Any non-substantial improvements made using the inventive concept and technical solution of this invention are within the protection scope of this invention.
Claims
1. A method for measuring the outline dimensions of a truck cargo box and correcting the loading and unloading position, characterized in that: Includes the following steps: (1) Select the location for measuring the outline of the truck body; (2) The outline of the truck body is repeatedly scanned by lidar to collect multiple frames of effective point cloud data; (3) Fusion of truck body point cloud contour data; (4) Straight-line fitting in the length and width directions of the carriage; (5) Obtain the outline line corresponding to the interior space of the carriage, and calculate the actual length and width of the carriage interior accordingly; The method further includes: Step (6) During the loading and unloading task of the robot, the area around the task target point of the robot loading and unloading is scanned in real time to determine whether there is a collision, and the task target point is corrected based on the collision. In step (4), the point cloud is fused based on a single frame. The minimum 2D bounding box (OBB) of the truck body contour map point cloud is calculated. Based on the major and minor axis poses of the OBB bounding box, the approximate pose of the truck body is determined. The major and minor axis poses are used as pose constraints for line fitting. The Random Sample Consensus (RANSAC) method is used to continuously perform parallel line fitting and Euclidean clustering on the truck body contour map point cloud to obtain the parallel line poses on both sides of the truck body and the vertical line pose of the truck body head. Step (5) includes: 1) After obtaining the pose parameters of the parallel straight lines on both sides of the truck bed and the pose parameters of the vertical straight line at the front of the truck bed, respectively, the three straight lines are... Move parallel to the inside of the truck bed to obtain three new straight lines. Until single-frame point cloud fusion All contour points corresponding to the three straight lines are located on the new three straight lines. The description of the outer outline of the carriage, at this point, shows three straight lines. The corresponding carriage size is taken as the effective size of the actual interior space of the carriage; 2) Set the new three straight lines of the truck bed. Parallel lines on both sides perpendicular to the front of the carriage The intersection points are respectively , ,use and Divide the outer contour area of the carriage into , and Three regions, moving the parallel lines on both sides away The area formed is , The region will Region and The corresponding carriage outline points in the region are respectively directed towards and Projection is performed, and the distance from the projection point is... , The farthest point and These are the two other points at the very edge of the rear of the car. and ; 3) Based on acquisition , , and Calculate the actual length of the carriage interior using four points. and width : ； ； In step (6), during the robot's loading and unloading task, a collision avoidance control zone is drawn in front of the robot. As the robot slowly approaches the placement point, the presence of the vehicle outline point cloud within the collision avoidance zone is monitored in real time. If the vehicle outline point cloud is found in the collision avoidance control zone in front of the robot, it is determined that there is a collision between the collision avoidance control zone in front of the robot and the vehicle outline. At this time, it is determined whether the actual point cloud outline around the cargo deployment point will collide with the cargo. If a collision occurs, a collision is initiated along a straight line perpendicular to the head of the vehicle. The direction is shifted inwards to adjust the loading and unloading position.
2. The method for measuring the outline dimensions of a truck cargo box and correcting the loading and unloading position as described in claim 1, characterized in that: Step (1) includes: parking the robot in the truck compartment contour measurement area according to the truck external robot navigation and positioning system, and ensuring that the scanning field of the LiDAR at the truck compartment contour measurement position where the robot is parked covers the interior area of the truck compartment.
3. The method for measuring the outline dimensions of a truck cargo box and correcting the loading and unloading position as described in claim 1, characterized in that: Step (2) includes: scanning the outline of the truck body multiple times using at least one lidar to obtain... Frame-by-frame carriage outline point cloud; point cloud preprocessing is performed on each frame of the multi-frame LiDAR point cloud data obtained by scanning to extract the effective carriage interior point cloud data of each frame.
4. The method for measuring the outline dimensions of a truck cargo box and correcting the loading and unloading position as described in claim 3, characterized in that: Each frame of the multi-frame LiDAR point cloud data obtained by scanning is preprocessed, including point cloud filtering, region growing, and contour interpolation. Point cloud filtering: for Each frame of the carriage outline point cloud is subjected to point cloud filtering processing, and each point in each frame of the point cloud is calculated. Distance to the origin of the point cloud coordinate system of this frame Based on the actual maximum possible effective distance for parking the truck bed. and the maximum effective value of radar ranging Filter out all unrealistic and unreasonable noise points in the point cloud; Region growth: Based on a fixed constraint angle of the radar scanned interior profile of the carriage. A small portion of the interior point cloud of the train carriage was directly extracted from the filtered point cloud. Then, by calculating the Euclidean distance between adjacent points, the small part of the point cloud is... Expand the regions to both sides to complete the point cloud of all the outlines inside the carriage; Contour interpolation: Point cloud representation of the complete interior contour of the carriage Perform interpolation on the point cloud According to laser frames Calculate the Euclidean distance between two points by the order of the points in the equation. Maximum constraint distance between selected points This yields the minimum number of complementary points between two points. And fill in evenly according to the minimum number of fill points. By analyzing the data from each point, we can obtain all valid point cloud data for each frame.
5. The method for measuring the outline dimensions of a truck cargo box and correcting the loading and unloading position as described in claim 1, characterized in that: Step (3) includes: The point cloud data is fused from all valid point cloud data extracted by multi-frame LiDAR and the robot pose corresponding to each valid data frame; each LiDAR laser frame... All correspond to the robot pose in that frame scan. , The secondary mapping scan will collect... For laser frames and robot pose Based on the robot's pose The point clouds of the next laser frame are unified into the same reference coordinate system. Next, the All points in the frame point cloud Unified to coordinate system The transformed point cloud is obtained below. : ； in ; It refers to the robot's pose relative to the reference coordinate system during each mapping process. The homogeneous transformation matrix; For unified to a reference coordinate system Grouped point clouds All points are scanned according to the laser frame sequence. All points in the same sequence are fused into a single-frame fused point cloud. : ； in, These are the coordinates of the point with index 'index' in the fused single-frame point cloud. It refers to the coordinates of the point with index 'index' in the single-frame point cloud before fusion.