Steel seat frame bayonet accurate positioning point cloud processing method and system based on grid point marking

Abstract

The invention provides a point cloud processing method and system for accurate positioning of a steel seat frame bayonet based on grid point marking. The method comprises the following steps: S1, acquiring point cloud data of a steel seat frame; s2, performing global coordinate conversion on the point cloud data of the steel seat frame by using calibration parameters of the 3D camera; s3, cleaning interference data of the point cloud data after coordinate conversion by using a noisy point filtering algorithm; s4, performing rasterization processing on the processed point cloud data to generate an ordered point cloud, and marking grids in a preset range of a gate; and S5, based on the marked grid segmentation bayonet point cloud data in the preset range of the bayonet, positioning the center coordinates of the bayonet.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a point cloud processing method and system for precise positioning of a steel frame bayonet based on grid point marking. Background Art

[0002] In industrial logistics, steel frames are used to secure steel coils during train loading. Because unmanned aerial vehicles (UAVs) are used in coil storage areas for lifting operations, including loading steel frames onto train cars, precise positioning of the steel frame's latch is crucial for ensuring the safety of the AUVs, accurate coil unloading, and overall coil transportation. While existing point cloud data-based positioning technology has made progress in areas such as autonomous driving and robotic navigation, precise positioning of the steel frame's latch has yet to be applied.

[0003] Patent document CN201951834U (application number: 201120019932.2) discloses a split steel seat frame for transporting coiled steel, which includes: left and right rectangular frame bases that cooperate with each other, the left rectangular frame base is fixedly connected to eight Λ-shaped support seats symmetrically distributed in two rows and four columns along the longitudinal and transverse center lines of the left rectangular frame base, and the right rectangular frame base is fixedly connected to eight Λ-shaped support seats symmetrically distributed in two rows and four columns along the longitudinal and transverse center lines of the right rectangular frame base, and base blocks are provided on the outer sides of the left and right rectangular frame bases. This patent focuses on the mechanical design of the steel seat frame. The present invention uses point cloud data to identify and locate the steel seat frame to guide the alignment of the steel seat frame and the train car wedge during the lifting process, which is inconsistent with the above-mentioned steel seat frame structure design direction. Summary of the Invention

[0004] In view of the defects in the prior art, the purpose of the present invention is to provide a point cloud processing method and system for precise positioning of the steel seat frame bayonet based on grid point marking.

[0005] According to the present invention, a point cloud processing method for precise positioning of a steel frame bayonet based on grid point marking is provided, comprising:

[0006] Step S1: Obtaining point cloud data of the steel frame;

[0007] Step S2: Using the 3D camera calibration parameters to perform global coordinate transformation on the steel frame point cloud data;

[0008] Step S3: Applying a noise filtering algorithm to the point cloud data after coordinate transformation to clean the interference data;

[0009] Step S4: rasterizing the processed point cloud data to generate an ordered point cloud, and marking the grids within the preset range of the checkpoint;

[0010] Step S5: segmenting the bayonet point cloud data based on the grid within the preset range of the marked bayonet, and locating the bayonet center coordinates.

[0011] Preferably, step S3 includes:

[0012] Step S3.1: pre-set the region of interest in the shooting field of view;

[0013]

[0014] Among them, P(x,y,z) represents the three-dimensional point in the region of interest;

[0015] Step S3.2: constructing a kdtree point cloud data structure based on the pre-set region of interest in the shooting field of view and establishing a neighborhood relationship between point clouds;

[0016] Set the outlier screening number thresh_count, mark the number of neighboring points of each point, and screen out isolated point information;

[0017]

[0018] Among them, real_count represents the actual number of neighborhood points of the current point, 1 is an isolated point mark, and 0 is a non-isolated point mark.

[0019] Preferably, the step S4 includes:

[0020] Step S4.1: Using the z-axis as the projection axis and the xy-plane as the projection surface, map the point cloud data to the xy-plane, with the z value being the current grid pixel value value_z_rgb, and create a plane mapping image of the starting point cloud;

[0021]

[0022] Among them, Xmax and Xmin represent the maximum and minimum values of the point cloud data on the X axis; img_length and img_width represent the length and width of the image; Ymax and Ymin represent the maximum and minimum values of the point cloud data on the Y axis; grid represents the set grid size;

[0023]

[0024] Among them, img row 、img col Respectively represent the row and column positions of the current data point pt in the image, pt x ,pt y Respectively represent the coordinates x and y of the data point pt;

[0025] Step 4.2: Mark pt(x,y) according to the pixel value value_z_rgb, initialize the marked value to 0 (no_find), and mark whether it is 1 according to its relationship with the set threshold thersh_rgb;

[0026]

[0027] Among them, pt(x,y) label represents the marked grid point pt, and 1 represents being marked.

[0028] Preferably, the step S5 includes:

[0029] Step S5.1: Extract the two-dimensional data point set pt Sum {pt1(x,y), pt2(x,y)…} of the edge where the bayonet is located. Combine the linear features of the edge, and perform least squares linear fitting on the extracted edge point cloud to generate the edge straight line equation Ax + By + C = 0;

[0030] Step S5.2: Use the generated edge straight line after fitting to segment the grid point cloud data of the steel seat frame, and extract the grid data point set protruding at the bayonet;

[0031] [[ID=​​​​​​​​​​​​​​​​​​​​​​​According to the present invention, a point cloud processing system for precise positioning of a steel frame bayonet based on grid point marking is provided, comprising:

[0038] Module M1: Obtain point cloud data of the steel frame;

[0039] Module M2: Use 3D camera calibration parameters to perform global coordinate transformation on the steel frame point cloud data;

[0040] Module M3: Apply noise filtering algorithm to the point cloud data after coordinate transformation to clean the interference data;

[0041] Module M4: rasterize the processed point cloud data to generate an ordered point cloud, and mark the grids within the preset range of the card slot;

[0042] Module M5: Segment the bayonet point cloud data based on the grid within the preset range of the marked bayonet and locate the bayonet center coordinates.

[0043] Preferably, the module M3 includes:

[0044] Module M3.1: Pre-set the area of interest in the shooting field of view;

[0045]

[0046] Where P(x,y,z) represents the three-dimensional point in the region of interest;

[0047] Module M3.2: Construct a kdtree point cloud data structure based on the pre-set region of interest in the shooting field of view and establish the neighborhood relationship between point clouds;

[0048] Set the outlier screening number thresh_count, mark the number of neighboring points of each point, and screen out isolated point information;

[0049]

[0050] Among them, real_count represents the actual number of neighborhood points of the current point, 1 is an isolated point mark, and 0 is a non-isolated point mark.

[0051] Preferably, the module M4 includes:

[0052] Module M4.1: Use the z-axis as the projection axis and the xy-plane as the projection surface to map the point cloud data to the xy-plane. The z-value is the current grid pixel value value_z_rgb, and a plane mapping image of the starting point cloud is created.

[0053]

[0054] Among them, Xmax and Xmin represent the maximum and minimum values of the point cloud data on the X axis; img_length and img_width represent the length and width of the image; Ymax and Ymin represent the maximum and minimum values of the point cloud data on the Y axis; grid represents the set grid size;

[0055]

[0056] Among them, img row 、img col Respectively represent the row and column positions of the current data point pt in the image, pt x ,pt y Respectively represent the coordinates x and y of the data point pt;

[0057] Step 4.2: Mark pt(x,y) according to the pixel value value_z_rgb, initialize the marker value to 0 (no_find), and mark it as 1 based on its relationship with the set threshold thersh_rgb;

[0058]

[0059] Among them, pt(x,y) label Indicates the marked grid point pt, 1 indicates it is marked.

[0060] Preferably, the module M5 includes:

[0061] Module M5.1: Extract the two-dimensional data point set pt at the edge of the card slot Sum {pt1(x,y),pt2(x,y)…}, combined with the straight line features at the edge, the extracted edge point cloud is subjected to least squares linear fitting processing to generate the edge line equation Ax+By+C=0;

[0062] Module M5.2: Use the edge lines generated after fitting to perform linear segmentation on the grid point cloud data of the steel frame and extract the grid data point set protruding at the bayonet;

[0063] Module M5.3: Apply k-means clustering to obtain raster data of n block-shaped protruding objects;

[0064] Module M5.4: The target bayonet is composed of the two block-shaped protrusions closest to the center of the mount. The n clustered protrusions are sorted according to their distance from the center of the mount, and the two closest ones are selected as the final bayonet protrusions.

[0065] Module M5.5: Calculate the coordinates pt of the plane midpoint of the line connecting the centers of two blocks center (x,y), coordinates of the midpoint of the plane pt center(x, y) is the coordinate pt of the target bayonet target (x, y).

[0066] Preferably, the module M5.2 includes:

[0067] Module M5.2.1: Calculate the distance Dist from the grid data point pt(x, y) to the straight line Ax + By + C = 0;

[0068] Module M5.2.2: Set the distance threshold D_Threshold, and retain the grid data with Dist < D_Threshold to form a data point set Sum{pt1, pt2,..., ptn}.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] 1. The present invention uses lidar point cloud data to solve the problem of high-precision positioning of the grasping points of the steel seat frame;

[0071] ]2. The present invention converts unstructured 3D data into structured 2D data through feature extraction processing based on grid processing, and solves the algorithm unity problem of bayonet positioning for steel seat frames of different specifications through dimensionality reduction and structured processing; More specifically, based on grid-based dimensionality reduction processing, the disordered point cloud is converted into an ordered point cloud for processing;

[0072] 3. The present invention performs bayonet clustering segmentation recognition and groove center positioning based on the protruding features of the steel seat frame bayonet. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] By reading the detailed description of the non-restrictive embodiments with reference to the following drawings, other features, objects and advantages of the present invention will become more apparent:

[0074] Figure 1 It is a flowchart of a point cloud processing method for precise positioning of a steel seat frame bayonet based on grid point marking.

[0075] Figure 2a , Figure 2b It is a schematic diagram of the alignment of the steel seat frame bayonet and the carriage wedge iron.

[0076] Figure 3 It is a schematic diagram of a steel seat frame photographed by a ToF camera.

[0077] Figure 4 It is a schematic diagram of fitting a straight line to the edge of the steel seat frame.

[0078] Figure 5 It is a schematic diagram of the bayonet. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0079] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0080] Example 1

[0081] According to the present invention, a point cloud processing method and system for precise positioning of a steel frame bayonet based on grid point marking are provided, comprising: obtaining point cloud data of the steel frame by installing a ToF camera on a special clamp for the steel frame.

[0082] The ToF camera is a 3D imaging device based on the time-of-flight principle. Before loading the steel coils onto the train, the steel frame needs to be hoisted into the train compartment in advance, and then the steel coils need to be hoisted onto the steel frame. The steel frame is the carrier for the steel coils to be placed in the train compartment. Before placing the steel coils in the train compartment, the steel frame's bayonet needs to be obtained so that it can be accurately aligned with the protruding wedge in the train compartment. Figure 2a 、 Figure 2b shown.

[0083] The point cloud processing method for accurate positioning of the steel frame bayonet based on grid point marking is as follows: Figure 1 Shown, including:

[0084] Step 1: Get the point cloud data of the steel frame, such as Figure 3 As shown;

[0085] Step 2: Use the 3D camera calibration parameters to perform global coordinate transformation on the steel frame point cloud data;

[0086] Step 3: Apply noise filtering algorithm to the point cloud data after coordinate transformation to clean the interference data;

[0087] Step 4: Rasterize the point cloud data to generate an ordered point cloud and mark the grids near the checkpoints;

[0088] Step 5: Apply feature extraction optimization algorithm to segment the precise card holder point cloud data and locate the card holder center coordinates (x, y, z).

[0089] Specifically, step 3 includes the following steps:

[0090] Step 3.1: Pre-set the area of interest in the shooting field of view to reduce the amount of point cloud data processing and improve data processing speed.

[0091]

[0092] Wherein, P(x, y, z) represents a three-dimensional point in space. In this embodiment, minX = -2500 mm, maxX = 2500 mm, minY = -1000 mm, maxY = 1000 mm, minZ = 1000 mm, and maxZ = 4500 mm correspond to the value ranges of the x, y, and z axes, respectively.

[0093] Step 3.2: Construct a kdtree point cloud data structure and establish neighborhood relationships between point clouds. Set the outlier screening number thresh_count (in this embodiment, the outlier screening number thresh_count = 20), mark the neighborhood point count (real_count) of each point, and screen out isolated point information.

[0094]

[0095] Among them, 1 is the isolated point mark and 0 is the non-isolated point mark.

[0096] Specifically, step 4 includes the following steps:

[0097] Step 4.1: Use the z-axis as the projection axis and the xy-plane as the projection surface to map the point cloud data to the xy-plane. The z value is the current grid pixel value value_z_rgb, and a plane mapping image of the starting point cloud is created.

[0098]

[0099] Among them, Xmax and Xmin represent the maximum and minimum values of the point cloud data on the X axis; img_length and img_width represent the length and width of the image; Ymax and Ymin represent the maximum and minimum values of the point cloud data on the Y axis; grid represents the set grid size.

[0100]

[0101] Among them, img row 、img col Respectively represent the row and column positions of the current data point pt in the image, pt x ,pt y Respectively represent the x and y coordinates of the data point pt.

[0102] Step 4.2: Mark pt(x, y) according to the pixel value value_z_rgb, initialize the mark value to 0 (no_find), and mark it as 1 (found) based on its relationship with the set threshold thersh_rgb. In this embodiment, thersh_rgb∈(160, 200).

[0103]

[0104] where pt(x,y) label Indicates the marked grid point pt, 1 indicates it is marked.

[0105] In order to accurately extract the straight line boundary of the steel frame bayonet in step 5 and reduce the influence of the steel frame hump on the recognition effect, it is necessary to cluster the point cloud near the steel frame bayonet. The specific clustering processing method is as follows:

[0106] 1) Randomly scan the image and find the pixel (x0, y0) with label 1 as the initial seed point, assign it the attribute found, and establish the 8-neighborhood point domain of this pixel;

[0107] 2) If the label of the neighborhood point is 1 and the attribute is no_find, mark it as a seed point, assign it the attribute found, and save the point; if the label of the neighborhood point is 0, skip the point and go to the next neighborhood;

[0108] 3) Traverse all seed points and perform the above neighborhood search operation until the seed point is empty;

[0109] 4) Traverse the remaining points in the image with label 1 and attribute no_find, repeat 2 and 3 until all points with label 1 are searched;

[0110] 5) In the segmented point cloud data, extract the data P based on the known plane size information of the steel frame set {pt1, pt2, …} is used in the subsequent step 5 to extract the steel seat frame bayonet contour information.

[0111] Specifically, step 5 includes the following steps:

[0112] Step 5.1: In order to accurately obtain the location information of the steel frame bayonet, it is necessary to further identify and extract the contour near the bayonet and extract the two-dimensional data point set pt at the edge of the bayonet. Sum {pt1(x,y),pt2(x,y)…}, combined with the straight line features at the edge, the extracted edge point cloud is subjected to least squares linear fitting processing ( Figure 4 Red marked straight line), generate edge straight line equation Ax+By+C=0;

[0113] Step 5.2: Furthermore, since the bayonet has a protruding feature at the edge of the steel frame, the steel frame grid point cloud data is segmented using a fitted two-dimensional straight line to extract the protruding grid data point set at the bayonet.

[0114] A) Calculate the distance Dist from the raster data point pt(x, y) to the line Ax + By + C = 0, which should be the distance on one side of the line;

[0115] B) Set the distance threshold D_Threshold = 10 mm, and retain the raster data with Dist < D_Threshold to form a data point set Sum{pt1, pt2,..., ptn};

[0116] Step 5.3: Further, since there are 4 blocky protrusions near the target bayonet, as Figure 5 shown, apply k-means clustering (where k = 4) to obtain the raster data of the 4 blocky protrusions.

[0117] Step 5.4: Further, the target bayonet is composed of the two blocky protrusions that are the closest to the center of the seat frame in the (x direction). Sort the 4 clustered protrusions according to their distances from the center of the seat frame in the (x direction), and select the 2 closest ones, which are the final bayonet protrusions.

[0118] Step 5.5: Calculate the coordinates pt center (x, y) of the midpoint of the line connecting the centers of the two blocky objects, which are the coordinates pt target (x, y) of the target bayonet.

[0119] The present invention also provides a point cloud processing system for precise positioning of the steel seat frame bayonet based on raster point marking. The point cloud processing system for precise positioning of the steel seat frame bayonet based on raster point marking can be implemented by executing the process steps of the point cloud processing method for precise positioning of the steel seat frame bayonet based on raster point marking. That is, those skilled in the art can understand the point cloud processing method for precise positioning of the steel seat frame bayonet based on raster point marking as the preferred implementation manner of the point cloud processing system for precise positioning of the steel seat frame bayonet based on raster point marking.

[0120] Those skilled in the art know that in addition to implementing the systems, devices, and their respective modules provided by the present invention in the form of pure computer-readable program code, the method steps can be logically programmed to enable the systems, devices, and their respective modules provided by the present invention to be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers, etc. to achieve the same program. Therefore, the systems, devices, and their respective modules provided by the present invention can be considered as a kind of hardware component, and the modules included therein for implementing various programs can also be regarded as the structures within the hardware component; the modules for implementing various functions can also be regarded as both software programs for implementing the method and the structures within the hardware component.

[0121] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.

Claims

1. A point cloud processing method for accurate positioning of steel frame bayonet based on grid point marking, characterized in that: including: Step S1: Obtain the point cloud data of the steel seat frame; Step S2: Perform global coordinate transformation on the point cloud data of the steel seat frame using the calibration parameters of the 3D camera; Step S3: Apply a noise filtering algorithm to the point cloud data after coordinate transformation to clean the interference data; Step S4: Perform rasterization processing on the processed point cloud data to generate an ordered point cloud, and mark the grids within the preset range of the bayonet; Step S5: Segment the bayonet point cloud data based on the marked grids within the preset range of the bayonet, and locate the center coordinates of the bayonet.

2. The point cloud processing method for accurate positioning of the steel frame bayonet based on grid point marking according to claim 1 is characterized in that: The said Step S3 includes: Step S3.1: Preset the region of interest in the shooting field of view; where, P(x, y, z) represents the three-dimensional spatial points of the region of interest; Step S3.2: Based on the preset region of interest in the shooting field of view, construct a kdtree point cloud data structure and establish the neighborhood relationship between point clouds; Set the outlier screening number thresh_count, mark the information of the number of neighborhood points of each point, and screen the outlier information; where, real_count represents the actual number of neighborhood points of the current point, 1 is the outlier mark, and 0 is the non-outlier mark.

3. The point cloud processing method for accurate positioning of the steel frame bayonet based on grid point marking according to claim 1 is characterized in that: The said Step S4 includes: Step S4.1: Take the z-axis as the projection axis and the xy plane as the projection plane, map the point cloud data to the xy plane, the z value is the current grid pixel value value_z_rgb, and establish the plane mapping image of the point cloud; Mapping image length and width information where, Xmax and Xmin represent the maximum and minimum values of the point cloud data on the X axis; img_length and img_width represent the length and width of the image; Ymax and Ymin represent the maximum and minimum values of the point cloud data on the Y axis; grid represents the set grid size; Grid point information Among them, img row 、img col Respectively represent the row and column positions of the current data point pt in the image, pt x ,pt y Respectively represent the coordinates x and y of the data point pt; Step 4.2: Mark pt(x, y) according to the pixel value value_z_rgb, initialize the mark value to 0 (no_find), and mark whether it is 1 according to its relationship with the set threshold thersh_rgb; Among them, pt(x,y) label Indicates the marked grid point pt, 1 indicates it is marked.

4. The point cloud processing method for accurate positioning of the steel frame bayonet based on grid point marking according to claim 1 is characterized in that: The said Step S5 includes: Step S5.1: Extract the two-dimensional data point set pt at the edge of the card slot Sum {pt1(x,y),pt2(x,y)…}, combined with the straight line features at the edge, the extracted edge point cloud is subjected to least squares linear fitting processing to generate the edge line equation Ax+By+C=0; Step S5.2: Use the edge straight line generated after fitting to perform straight line segmentation on the grid point cloud data of the steel seat frame, and extract the grid data point set protruding at the bayonet; Step S5.3: Apply k-means clustering processing to obtain the grid data of n块状凸出物体 (blocky protruding objects); Step S5.4: The target bayonet is composed of the two blocky protrusions closest to the center of the seat frame. Sort the n protrusions clustered according to the distance from the center of the seat frame, and screen the 2 closest ones as the final bayonet protrusions; Step S5.5: Calculate the coordinates pt of the plane midpoint of the line connecting the centers of the two blocks center (x,y), coordinates of the midpoint of the plane pt center (x,y) is the coordinate pt of the target mount target (x,y).

5. The point cloud processing method for accurate positioning of the steel frame bayonet based on grid point marking according to claim 4 is characterized in that: The said Step S5.2 includes: Step S5.2.1: Calculate the distance Dist from the grid data point pt(x, y) to the straight line Ax + By + C = 0; Step S5.2.2: Set the distance threshold D_Threshold, retain the grid data with Dist < D_Threshold, and form the data point set Sum{pt1, pt2,..., ptn}.

6. A point cloud processing system for precise positioning of steel frame bayonet based on grid point marking, characterized in that: including: [[ID= ​ Module M3: Apply a noise filtering algorithm to the point cloud data after coordinate transformation to clean up interference data; Module M4: Perform rasterization on the processed point cloud data to generate an ordered point cloud, and mark the grids within the preset range of the bayonet; Module M5: Segment the bayonet point cloud data based on the marked grids within the preset range of the bayonet to locate the center coordinates of the bayonet.

7. The point cloud processing system for precise positioning of steel frame bayonet based on grid point marking according to claim 6 is characterized in that: The module M3 includes: Module M3.1: Preset the region of interest in the shooting field of view; Among them, P(x, y, z) represents the three-dimensional spatial points of the region of interest; Module M3.2: Construct a kdtree point cloud data structure based on the preset region of interest in the shooting field of view, and establish the neighborhood relationship between point clouds; Set the outlier screening number thresh_count, mark the neighborhood point number information of each point, and screen out the isolated point information; Among them, real_count represents the actual neighborhood point number of the current point, 1 is the isolated point mark, and 0 is the non-isolated point mark.

8. The point cloud processing system for precise positioning of steel frame bayonet based on grid point marking according to claim 6 is characterized in that: The module M4 includes: Module M4.1: Take the z-axis as the projection axis and the xy plane as the projection plane, map the point cloud data to the xy plane, the z value is the current grid pixel value value_z_rgb, and establish a plane mapping image of the point cloud; Mapping image length and width information Among them, Xmax and Xmin represent the maximum and minimum values of the point cloud data on the X axis; img_length and img_width represent the length and width of the image; Ymax and Ymin represent the maximum and minimum values of the point cloud data on the Y axis; grid represents the set grid size; Grid point information Among them, img row 、img col Respectively represent the row and column positions of the current data point pt in the image, pt x ,pt y Respectively represent the coordinates x and y of the data point pt; Step 4.2: Mark pt(x, y) according to the pixel value value_z_rgb, initialize the mark value to 0 (no_find), and mark whether it is 1 according to its relationship with the set threshold thersh_rgb; Among them, pt(x,y) label Indicates the marked grid point pt, 1 indicates it is marked.

9. The point cloud processing system for precise positioning of steel frame bayonet based on grid point marking according to claim 6, characterized in that: The module M5 includes: Module M5.1: Extract the two-dimensional data point set pt at the edge of the card slot Sum {pt1(x,y),pt2(x,y)…}, combined with the straight line features at the edge, the extracted edge point cloud is subjected to least squares linear fitting processing to generate the edge line equation Ax+By+C=0; Module M5.2: Use the edge straight line generated after fitting to segment the grid point cloud data of the steel seat frame, and extract the grid data point set protruding at the bayonet; Module M5.3: Apply k-means clustering processing to obtain the grid data of n块状凸出物体 (blocky protruding objects); Module M5.4: The target bayonet is composed of the two blocky protrusions closest to the center of the seat frame. Sort the n protrusions clustered according to the distance from the center of the seat frame, and screen out the 2 closest ones as the final bayonet protrusions; Module M5.5: Calculate the coordinates pt of the plane midpoint of the line connecting the centers of two blocks center (x,y), coordinates of the midpoint of the plane pt center (x,y) is the coordinate pt of the target mount target (x,y).

10. The point cloud processing system for precise positioning of steel frame bayonet based on grid point marking according to claim 9, characterized in that: The module M5.2 includes: [[ID=...]]Module M5.2.1: Calculate the distance Dist from the grid data point pt(x, y) to the straight line Ax + By + C = 0; Module M5.2.2: Set the distance threshold D_Threshold, retain the grid data with Dist < D_Threshold, and form a data point set Sum{pt1, pt2,..., ptn}. It should be noted that the term "块状凸出物体" in the original text seems a bit unclear. I've just translated it as "blocky protruding objects" for the time. You may need to clarify this term for a more accurate translation.

Images