Space intelligent grasping pose estimation error conduction blocking method

Abstract

This invention discloses a method for blocking the propagation of attitude estimation errors in intelligent spatial grasping, relating to the field of intelligent grasping. It solves the problem of poor grasping accuracy in existing methods for blocking the propagation of attitude estimation errors in intelligent spatial grasping. The method includes the following steps: Step S1: Acquiring the spatial objects to be grasped within the real-time grasping space to obtain multiple pre-grasping spatial objects; collecting and analyzing laser sampling points on the pre-grasping objects; obtaining the attitude uncertainty corresponding to each pre-grasping object based on the analysis results to obtain spatial attitude acquisition data; Step S2: Screening the pre-selected spatial objects according to the spatial attitude acquisition data to obtain spatial object type screening data; Step S3: Intelligently grasping the graspable spatial objects according to the spatial object type screening data. This invention can improve the accuracy of intelligent grasping of spatial objects.

Description

Technical Field

[0001] This invention belongs to the field of intelligent grasping and relates to sensor technology. Specifically, it is a method for blocking the propagation of attitude estimation errors in spatial intelligent grasping. Background Technology

[0002] Current methods for blocking the propagation of attitude estimation errors have the following drawbacks when grasping spatial volumes: 1. Existing methods for blocking the propagation of grasping posture estimation errors are difficult to combine with the occlusion of the environment in which the grasping object is located, so as to distinguish between graspable and ungraspable objects. Under partial occlusion conditions, even if the object is occluded by surrounding objects, it is objectively difficult to carry out grasping operations directly, which increases the potential risk of terminal collision and grasping failure to a certain extent. 2. Existing methods for blocking the propagation of grasping posture estimation errors generally cannot complete the screening of grasping planes for the identified graspable space. Most of them directly use the preset default grasping planes for subsequent grasping planning, resulting in low grasping accuracy and easy occurrence of gripping accidents.

[0003] Therefore, this application proposes a method for blocking the propagation of spatial intelligent grasping attitude estimation errors. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method for blocking the propagation of attitude estimation errors in space intelligent grasping, thereby improving the accuracy of intelligent grasping of space objects.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for blocking the propagation of spatial intelligent grasping attitude estimation errors, comprising the following steps: Step S1: Acquire the spatial objects to be grasped in the real-time grasping space to obtain multiple pre-grab spatial objects. Collect and analyze the laser sampling points of the pre-grab objects. Based on the analysis results, obtain the attitude uncertainty corresponding to each pre-grab object to obtain spatial attitude acquisition data. Step S2: Acquire spatial attitude data, perform type screening on the pre-selected spatial objects, and obtain spatial object type screening data; Step S3: Intelligently capture the graspable spatial objects based on the spatial object type screening data.

[0006] Furthermore, in step S1, the specific steps are as follows: Step S11: Acquire the scene space where the capture terminal is located to obtain the real-time capture space, collect the space volume in the current scene space that needs to be captured to obtain multiple pre-capture space volumes; Step S12: Select any P1 space volume from the pre-selected space volumes and perform scene pose analysis on the P1 space volume. Based on the analysis results, obtain the pose uncertainty corresponding to the P1 space volume. Step S13: Obtain the attitude uncertainty corresponding to each pre-selected spatial volume to obtain spatial attitude acquisition data.

[0007] Furthermore, in step S12, the specific steps are as follows: Step S121: Acquire a top-view image of the P1 spatial volume to obtain a top-view image of the P1 space. Mark the image area covered by the P1 spatial volume on the top-view image of the P1 space to obtain the P1 entity image area. Acquire the geometric center point of the P1 entity image area to obtain the top-view center point of P1. Step S122: In the top view image of P1, draw two perpendicular lines through the top view center point of P1 to obtain the first central perpendicular line and the second central perpendicular line. Use the top view center point of P1 to divide the first central perpendicular line into the first sub-segment perpendicular line and the second sub-segment perpendicular line. Use the top view center point of P1 to divide the second central perpendicular line into the third sub-segment perpendicular line and the fourth sub-segment perpendicular line. Step S123: Perform occlusion analysis on the P1 space volume based on the first sub-segment vertical line, and obtain the entity occlusion degree corresponding to the first sub-segment vertical line according to the analysis results; Step S124: Obtain the solid obstruction degree corresponding to the vertical line from the second sub-segment to the fourth sub-segment, and calculate the average of the multiple solid obstruction degrees to obtain the solid average obstruction degree. Step S125: Create a reference value for the entity obstruction degree, calculate the difference between the reference value for the entity obstruction degree and the average obstruction degree of the entity, and obtain the attitude uncertainty corresponding to the P1 space volume.

[0008] Furthermore, in step S123, the specific steps are as follows: Edge contour extraction is performed on the P1 entity image region to obtain the P1 entity contour. The intersection point of the first sub-segment perpendicular line and the P1 entity contour is collected to obtain the first contour perpendicular line intersection point. The distance between the intersection point of the top-view center point of P1 and the first contour perpendicular line is collected to obtain the first vertical contour distance. The first sub-segment vertical line is discretized into several vertical line pixels. The distance between each vertical line pixel and the top-view center point of P1 is obtained, resulting in multiple vertical line pixel distances. Obtain the necessary lateral grasping distance corresponding to the grasping terminal, calculate the sum of the necessary lateral grasping distance and the first vertical profile distance to obtain the necessary center distance, and arbitrarily select a pre-selected survey pixel from the vertical pixel points whose vertical pixel distance is greater than the necessary center distance.

[0009] Furthermore, in step S123, the specific steps are as follows: Map the intersection of the first contour perpendicular lines to the actual physical space area where the P1 space body is located to obtain the first actual mapping point. Map the pre-selected survey pixel points to the actual physical space area where the P1 space body is located to obtain the second actual mapping point. Map the first sub-segment perpendicular line to the actual physical space area where the P1 space body is located to obtain the first actual perpendicular line. Draw a plane perpendicular to the first actual vertical line through the first actual mapping point to obtain the first pre-selected mapping plane. Obtain the vertical projection of the P1 spatial volume in the first pre-selected mapping plane to obtain the P1 spatial volume projection. Mark the plane area covered by the P1 spatial volume projection in the first pre-selected mapping plane as the pre-selected laser sampling area.

[0010] Furthermore, in step S123, the specific steps are as follows: The laser emission point corresponding to the laser emission terminal is controlled to coincide with the second actual mapping point, and the pre-selected laser sampling area is scanned to obtain multiple laser sampling points; Arbitrarily select a T1 laser sampling point from the acquired laser sampling points, perform linear fitting on the laser path corresponding to the T1 laser sampling point to obtain the T1 sampling line segment, and numerically obtain the angle between the T1 sampling line segment and the first actual perpendicular line at the second actual mapping point to obtain the T1 laser angle. Obtain the distance between the first and second actual mapping points, calculate the ratio of the obtained distance to the cosine of the angle between the T1 laser points, and obtain the T1 laser sampling reference path distance.

[0011] Furthermore, in step S123, the specific steps are as follows: The distance between the T1 laser sampling point and the second actual mapping point is obtained to obtain the actual path distance of the T1 laser. If the actual path distance of the T1 laser is less than the T1 laser sampling reference path distance, the T1 laser sampling point is screened into a path obstruction sampling point. If the actual path distance of the T1 laser is equal to the T1 laser sampling reference path distance, the T1 laser sampling point is screened into a space volume sampling point. Each laser sampling point is screened separately, and the ratio of the number of path-blocking sampling points to the number of laser sampling points is calculated to obtain the entity occlusion degree corresponding to the vertical line of the first sub-segment.

[0012] Furthermore, in step S2, the specific steps are as follows: Obtain space body type screening data, and obtain the attitude uncertainty corresponding to each pre-selected space body based on the space body type screening data; Create a preset feature value for uncertainty. If the attitude uncertainty is greater than the preset feature value, the pre-selected spatial volume is marked as an occluded spatial volume. If the attitude uncertainty is less than or equal to the preset feature value, the pre-selected spatial volume is marked as a graspable spatial volume, thus obtaining spatial volume type screening data.

[0013] Furthermore, step S3 also includes the following steps: Step S31: Obtain space body type screening data, and obtain the type of each pre-selected space body based on the space body type screening data; Step S32: Acquire the laser sampling points contained in each pre-selected spatial volume to obtain multiple laser point cloud sets; Step S33: If the pre-selected space is determined to be an occluded space, the transmission of the laser point cloud set to the grasping terminal is interrupted, and an occlusion removal operation is performed on the pre-selected space. After the occlusion removal is completed, the type screening process is carried out again until the pre-selected space is determined to be a graspable space. Step S34: If the pre-selected spatial object is determined to be a graspable spatial object, the laser point cloud set is transmitted to the grasping terminal. The grasping terminal performs appropriate grasping side combination filtering on the graspable spatial object and grasps the graspable spatial object through appropriate grasping side combination.

[0014] Furthermore, step S34 also includes the following steps: The spatial sides contained in the graspable space are acquired to obtain multiple graspable sides, and any two graspable sides that are in a parallel state are marked as a graspable side combination. Each grasping side combination is labeled as the first grasping side and the second grasping side. The number of sampling points in the space covered by the first grasping side is counted to obtain the number of samples of the first side. The area value of the first grasping side is obtained. The ratio of the number of samples of the first side to the area value of the first side is calculated to obtain the sampling ratio of the first side. Numerical analysis of spatial volume sampling points is performed on the second grasping side. The sampling ratio of the second side is obtained based on the analysis results. The sampling ratio of the first side and the sampling ratio of the second side are summed to obtain the total sampling ratio of the sides corresponding to each grasping side combination. The grasping side combination corresponding to the largest total sampling ratio is selected as the suitable grasping side combination.

[0015] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention makes graspability judgment by combining the occlusion state of the environment in which the space object is located. It can comprehensively consider side occlusion, top occlusion and surrounding objects. Instead of relying solely on the posture estimation output to directly enter the grasping process, it can identify potential grasping risks caused by occlusion in advance to a certain extent, and help improve the stability of the overall grasping task. 2. Based on the selection of graspable spatial objects, this invention adds a grasping plane selection step, which can select grasping operation planes with higher sampling density and better fit from multiple candidate grasping planes. This provides a more reasonable attitude basis for subsequent grasping planning, avoids attitude deviation caused by directly using the default plane, reduces the possibility of error propagation to subsequent steps from the source, and effectively improves the accuracy of grasping attitude estimation results. Attached Figure Description

[0016] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0017] Figure 1 This is a diagram illustrating the implementation steps of the present invention; Figure 2 This is a top-view diagram of the center point of P1 in this invention; Figure 3 This is a schematic diagram of the T1 laser angle of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1 Please see Figure 1 This invention provides a technical solution: a method for blocking the propagation of attitude estimation errors in spatial intelligent grasping, comprising the following steps: Step S1: Acquire the spatial objects to be grasped in the real-time grasping space to obtain multiple pre-grab spatial objects. Collect and analyze the laser sampling points of the pre-grab objects. Based on the analysis results, obtain the attitude uncertainty corresponding to each pre-grab object to obtain spatial attitude acquisition data. In step S1, the specific steps are as follows: The scene space where the capture terminal is located is acquired to obtain the real-time capture space. The space volume that needs to be captured in the current scene space is collected to obtain multiple pre-capture space volumes.

[0020] It should be noted here that: In this application, the real-time grasping space referred to herein specifically refers to the complete three-dimensional working environment in which the grasping terminal is located, and the grasping terminal referred to herein is a grasping robot; In this application, the spatial body referred to herein is specifically a three-dimensional object entity in the current scene space that has a three-dimensional solid form and can be grasped and manipulated by a robot.

[0021] Arbitrarily select a P1 space volume from the obtained pre-selected space volumes, and perform scene pose analysis on the P1 space volume. Based on the analysis results, obtain the pose uncertainty corresponding to the P1 space volume. Specifically as follows: A top-view image of the P1 spatial volume is acquired. The image area covered by the P1 spatial volume is marked in the top-view image of the P1 spatial volume to obtain the P1 entity image area. The geometric center point of the P1 entity image area is acquired to obtain the top-view center point of P1. It should be noted here that: In this application, the P1 space volume referred to herein is a cuboid that is occluded by objects on all sides but not by objects on top.

[0022] Please see Figure 2 In the top-view image of space P1, draw two perpendicular lines through the top-view center point of P1 to obtain the first central perpendicular line and the second central perpendicular line. Use the top-view center point of P1 to divide the first central perpendicular line into the first sub-segment perpendicular line and the second sub-segment perpendicular line. Use the top-view center point of P1 to divide the second central perpendicular line into the third sub-segment perpendicular line and the fourth sub-segment perpendicular line. Based on the first segment perpendicular line, an occlusion analysis is performed on the P1 space volume, and the occlusion degree of the entity corresponding to the first segment perpendicular line is obtained according to the analysis results. Specifically as follows: Edge contour extraction is performed on the P1 entity image region to obtain the P1 entity contour. The intersection point of the first sub-segment perpendicular line and the P1 entity contour is collected to obtain the first contour perpendicular line intersection point. The distance between the intersection point of the top-view center point of P1 and the first contour perpendicular line is collected to obtain the first vertical contour distance. The first sub-segment vertical line is discretized into several vertical line pixels. The distance between each vertical line pixel and the top-view center point of P1 is obtained, resulting in multiple vertical line pixel distances. Obtain the necessary lateral grasping distance corresponding to the grasping terminal, calculate the sum of the necessary lateral grasping distance and the first vertical profile distance to obtain the necessary center distance, and arbitrarily select a pre-selected survey pixel from the vertical pixel points whose vertical pixel distance is greater than the necessary center distance. It should be noted here that: In this application, the necessary lateral gripping spacing referred to herein is specifically the minimum three-dimensional spatial gap that must be reserved on the lateral periphery of the P1 space body in order to complete the lateral gripping.

[0023] Map the intersection of the first contour perpendicular lines to the actual physical space area where the P1 space volume is located to obtain the first actual mapping point. Map the pre-selected survey pixel points to the actual physical space area where the P1 space volume is located to obtain the second actual mapping point. Map the first sub-segment perpendicular line to the actual physical space area where the P1 space volume is located to obtain the first actual perpendicular line.

[0024] Draw a plane perpendicular to the first actual vertical line through the first actual mapping point to obtain the first pre-selected mapping plane. Obtain the vertical projection of the P1 spatial volume in the first pre-selected mapping plane to obtain the P1 spatial volume projection. Mark the plane area covered by the P1 spatial volume projection in the first pre-selected mapping plane as the pre-selected laser sampling area. The laser emission point corresponding to the laser emission terminal is controlled to coincide with the second actual mapping point, and the pre-selected laser sampling area is scanned to obtain multiple laser sampling points; It should be noted here that: In this application, the laser emitting terminal referred to herein is a lidar.

[0025] Please see Figure 3 Arbitrarily select a T1 laser sampling point from the acquired laser sampling points, perform linear fitting on the laser path corresponding to the T1 laser sampling point to obtain the T1 sampling line segment, and numerically obtain the angle between the T1 sampling line segment and the first actual perpendicular line at the second actual mapping point to obtain the T1 laser angle. Obtain the distance between the first actual mapping point and the second actual mapping point, calculate the ratio of the obtained distance to the cosine of the angle between the two points, and obtain the sampling reference path distance of the T1 laser. The distance between the T1 laser sampling point and the second actual mapping point is obtained to obtain the actual path distance of the T1 laser. If the actual path distance of the T1 laser is less than the T1 laser sampling reference path distance, the T1 laser sampling point is screened into a path obstruction sampling point. If the actual path distance of the T1 laser is equal to the T1 laser sampling reference path distance, the T1 laser sampling point is screened into a space volume sampling point. It should be noted here that: In this application, since the laser beam cannot penetrate the interior of the P1 space, the actual path distance of the T1 laser is always less than or equal to the laser sampling reference path distance.

[0026] Repeat the screening process for the T1 laser sampling point, screen each laser sampling point separately, and calculate the ratio of the number of path blocking sampling points to the number of laser sampling points to obtain the entity occlusion degree corresponding to the first sub-segment vertical line. Repeat the process of obtaining the entity occlusion degree corresponding to the first sub-segment vertical line, obtain the entity blocking degree corresponding to the second to fourth sub-segment vertical lines respectively, and calculate the average of the obtained entity blocking degree to obtain the entity average blocking degree.

[0027] Create a reference value for the entity's obstruction, calculate the difference between the reference value and the average obstruction of the entity, and obtain the attitude uncertainty corresponding to the P1 space volume. It should be noted here that: In this application, the reference value for the physical obstruction degree is "1".

[0028] Repeat the process of acquiring the attitude uncertainty corresponding to the P1 spatial volume, and acquire the attitude uncertainty corresponding to each pre-selected spatial volume to obtain spatial attitude acquisition data. Step S2: Acquire spatial attitude acquisition data, and perform type screening on the pre-selected spatial objects based on the spatial attitude acquisition data to obtain spatial object type screening data; In step S2, the specific steps are as follows: Obtain space body type screening data, and obtain the attitude uncertainty corresponding to each pre-selected space body based on the space body type screening data; Create a preset feature value for uncertainty. If the attitude uncertainty is greater than the preset feature value, the pre-selected spatial volume is marked as an occluded spatial volume. If the attitude uncertainty is less than or equal to the preset feature value, the pre-selected spatial volume is marked as a graspable spatial volume, thus obtaining spatial volume type screening data.

[0029] It should be noted here that: In this application, the attitude uncertainty of historical spatial objects that have been identified as graspable spatial objects is calculated to obtain multiple historical attitude uncertainties. The obtained multiple historical attitude uncertainties are numerically compared, and the maximum historical attitude uncertainty is set as the uncertainty preset characteristic value.

[0030] Step S3: Intelligently capture graspable spatial objects based on the spatial object type screening data; Step S3 further includes the following steps: Obtain space body type screening data, and determine the type of each pre-selected space body based on the space body type screening data; The laser sampling points contained in each pre-selected spatial volume are acquired to obtain multiple laser point cloud sets; If the pre-selected space is determined to be an obstructing space, the transmission of the laser point cloud set to the grasping terminal is interrupted, and an obstruction removal operation is performed on the pre-selected space. After the obstruction removal is completed, the type screening process is carried out again until the pre-selected space is determined to be a graspable space. If the pre-selected spatial object is determined to be a graspable spatial object, the laser point cloud set is transmitted to the grasping terminal. The grasping terminal performs appropriate grasping side combination filtering on the graspable spatial object, and grasps the graspable spatial object through appropriate grasping side combination.

[0031] It should be noted here that: In this application, the grasping terminal referred to herein is a grasping robot.

[0032] Specifically as follows: The spatial sides contained in the graspable space are acquired to obtain multiple graspable sides, and any two graspable sides that are in a parallel state are marked as a graspable side combination. Each grasping side combination is labeled as the first grasping side and the second grasping side. The number of sampling points in the space covered by the first grasping side is counted to obtain the number of samples of the first side. The area value of the first grasping side is obtained. The ratio of the number of samples of the first side to the area value of the first side is calculated to obtain the sampling ratio of the first side. Numerical analysis of spatial volume sampling points is performed on the second grasping side. The sampling ratio of the second side is obtained based on the analysis results. The sampling ratio of the first side and the sampling ratio of the second side are summed to obtain the total sampling ratio of the sides corresponding to each grasping side combination. The grasping side combination corresponding to the largest total sampling ratio is selected as the suitable grasping side combination.

[0033] Compared to the problems described in the background art, the present invention makes graspability judgment by combining the occlusion state of the environment in which the spatial object is located. It can comprehensively consider side occlusion, top occlusion and the surrounding objects, and no longer directly enter the grasping process based solely on the attitude estimation output. It can identify potential grasping risks caused by occlusion in advance to a certain extent, and help improve the stability of the overall grasping task.

[0034] Furthermore, this invention adds a grasping plane screening step on the basis of screening the graspable space volume. It can screen the grasping operation plane with higher sampling density and better fit from multiple candidate grasping planes, providing a more reasonable attitude basis for subsequent grasping planning, avoiding the attitude deviation caused by directly using the default plane, reducing the possibility of error propagation to subsequent steps from the source, and effectively improving the accuracy of grasping attitude estimation results.

[0035] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementation methods. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The present invention is limited only to the technical solutions and their full scope and equivalents.

Claims

1. A method for blocking the propagation of attitude estimation errors in spatial intelligent grasping, characterized in that, include: Step S1: Acquire the spatial objects to be captured in the scene space to obtain multiple pre-captured spatial objects. Collect and analyze the laser sampling points of the pre-captured objects to obtain the attitude uncertainty corresponding to each pre-captured object and obtain the spatial attitude acquisition data. Step S2: Acquire spatial attitude acquisition data, and perform type screening on the pre-selected spatial objects based on the spatial attitude acquisition data to obtain spatial object type screening data; Step S3: Intelligently capture the graspable spatial objects based on the spatial object type screening data.

2. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 1, characterized in that, In step S1, the specific steps are as follows: Step S11: Acquire the scene space where the capture terminal is located to obtain the real-time capture space, collect the space volume in the current scene space that needs to be captured to obtain multiple pre-capture space volumes; Step S12: Select any P1 space volume from the acquired pre-selected space volumes, perform scene pose analysis on the P1 space volume, and obtain the pose uncertainty based on the analysis results; Step S13: Obtain the attitude uncertainty corresponding to each pre-selected spatial volume to obtain spatial attitude acquisition data.

3. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 2, characterized in that, In step S12, the specific steps are as follows: Step S121: Acquire a top-view image of the P1 spatial volume to obtain a top-view image of the P1 space. Mark the image area covered by the P1 spatial volume in the top-view image of the P1 space to obtain the P1 entity image area. Acquire the geometric center point of the P1 entity image area to obtain the top-view center point of P1. Step S122: In the top view image of P1, draw two perpendicular lines through the top view center point of P1 to obtain the first central perpendicular line and the second central perpendicular line. Use the top view center point of P1 to divide the central perpendicular line into sub-segment perpendicular lines. Step S123: Select the first segment vertical line from the segmented vertical lines, perform occlusion analysis on the P1 space volume based on the first segment vertical line, and obtain the entity occlusion degree corresponding to the first segment vertical line. Step S124: Obtain the solid obstruction degree corresponding to each segment vertical line, and calculate the average of the obtained solid obstruction degrees to obtain the solid average obstruction degree. Step S125: Create a reference value for the entity obstruction degree, calculate the difference between the reference value for the entity obstruction degree and the average obstruction degree of the entity, and obtain the attitude uncertainty corresponding to the P1 space volume.

4. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 3, characterized in that, In step S123, the specific steps are as follows: Edge contour extraction is performed on the P1 entity image region to obtain the P1 entity contour. The intersection point of the first sub-segment perpendicular line and the P1 entity contour is collected to obtain the first contour perpendicular line intersection point. The distance between the intersection point of the top-view center point of P1 and the first contour perpendicular line is collected to obtain the first vertical contour distance. Discretize the first sub-segment vertical line into several vertical line pixels, obtain the line distance between each vertical line pixel and the top-view center point of P1, and obtain multiple vertical line pixel distances; Obtain the necessary lateral gripping distance, calculate the sum of the necessary lateral gripping distance and the first vertical profile distance to obtain the necessary center distance, and arbitrarily select a pre-selected survey pixel from the vertical pixel points whose vertical pixel distance is greater than the necessary center distance.

5. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 4, characterized in that, In step S123, the specific steps are as follows: Map the intersection of the first contour perpendicular lines to the actual physical space area where the P1 space body is located to obtain the first actual mapping point. Map the pre-selected survey pixel points to the actual physical space area where the P1 space body is located to obtain the second actual mapping point. Map the first sub-segment perpendicular line to the actual physical space area where the P1 space body is located to obtain the first actual perpendicular line. Draw a plane perpendicular to the first actual vertical line through the first actual mapping point to obtain the first pre-selected mapping plane. Obtain the vertical projection of the P1 spatial volume in the first pre-selected mapping plane to obtain the P1 spatial volume projection. Mark the plane area covered by the P1 spatial volume projection in the first pre-selected mapping plane as the pre-selected laser sampling area.

6. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 5, characterized in that, In step S123, the specific steps are as follows: The laser emission point corresponding to the laser emission terminal is controlled to coincide with the second actual mapping point, and the pre-selected laser sampling area is scanned to obtain multiple laser sampling points; Arbitrarily select a T1 laser sampling point from the acquired laser sampling points, perform linear fitting on the laser path corresponding to the T1 laser sampling point to obtain the T1 sampling line segment, and numerically obtain the angle between the T1 sampling line segment and the first actual perpendicular line at the second actual mapping point to obtain the T1 laser angle. Obtain the distance between the first and second actual mapping points, calculate the ratio of the obtained distance to the cosine of the angle between the T1 laser points, and obtain the T1 laser sampling reference path distance.

7. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 6, characterized in that, In step S123, the specific steps are as follows: The distance between the T1 laser sampling point and the second actual mapping point is obtained to obtain the actual path distance of the T1 laser. If the actual path distance of the T1 laser is less than the T1 laser sampling reference path distance, the T1 laser sampling point is screened into a path obstruction sampling point. If the actual path distance of the T1 laser is equal to the T1 laser sampling reference path distance, the T1 laser sampling point is screened into a space volume sampling point. Each laser sampling point is screened, and the ratio of the number of path-blocking sampling points to the number of laser sampling points is calculated to obtain the entity occlusion degree corresponding to the vertical line of the first sub-segment.

8. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 1, characterized in that, In step S2, the specific steps are as follows: Obtain space body type screening data, and obtain the attitude uncertainty corresponding to each pre-selected space body based on the space body type screening data; Create a preset feature value for uncertainty. If the attitude uncertainty is greater than the preset feature value, the pre-selected spatial volume is marked as an occluded spatial volume. If the attitude uncertainty is less than or equal to the preset feature value, the pre-selected spatial volume is marked as a graspable spatial volume, thus obtaining spatial volume type screening data.

9. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 1, characterized in that, Step S3 further includes the following steps: Step S31: Obtain space body type screening data, and obtain the type of each pre-selected space body based on the space body type screening data; Step S32: Acquire the laser sampling points contained in each pre-selected spatial volume to obtain multiple laser point cloud sets; Step S33: If the pre-selected space is determined to be an occluded space, the transmission of the laser point cloud set to the grasping terminal is interrupted, and an occlusion removal operation is performed on the pre-selected space. After the occlusion removal is completed, the type screening process is carried out again until the pre-selected space is determined to be a graspable space. Step S34: If the pre-selected spatial object is determined to be a graspable spatial object, the laser point cloud set is transmitted to the grasping terminal. The grasping terminal performs appropriate grasping side combination filtering on the graspable spatial object and grasps the graspable spatial object through appropriate grasping side combination.

10. The spatial intelligent grasping attitude estimation error propagation blocking method according to claim 9, characterized in that, Step S34 further includes the following steps: The spatial sides contained in the graspable space are acquired to obtain multiple graspable sides, and any two graspable sides that are in a parallel state are marked as a graspable side combination. Each grasping side combination is labeled as the first grasping side and the second grasping side. The number of sampling points in the space covered by the first grasping side is counted to obtain the number of samples of the first side. The area value of the first grasping side is obtained. The ratio of the number of samples of the first side to the area value of the first side is calculated to obtain the sampling ratio of the first side. Numerical analysis of spatial volume sampling points is performed on the second grasping side to obtain the sampling ratio of the second side. The sampling ratio of the first side and the sampling ratio of the second side are summed to obtain the total sampling ratio of the sides corresponding to each grasping side combination. The grasping side combination corresponding to the maximum total sampling ratio of the sides is selected as the suitable grasping side combination.

Images