Grasp pose detection method, system, apparatus, robot, readable storage medium and program product

By extracting and fitting the first and second position data from the position data of the target box, the target plane is determined and the first contour graphic is translated, which solves the problem of inaccurate robot grasping posture estimation and improves the grasping success rate and stability.

CN122125702APending Publication Date: 2026-06-02CHONGQING PHOENIX TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHONGQING PHOENIX TECHNOLOGY CO LTD
Filing Date
2026-03-24
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

In existing technologies, the position data of the target box may be obstructed, noisy, or incomplete, which leads to inaccurate estimation of the robot's grasping posture of the target box and a low grasping success rate.

Method used

By extracting first position data (distance from the first surface within a first distance threshold range) and second position data (distance from a greater distance) from the target box's position data, shape fitting and principal component analysis are performed to determine the target plane. The first contour graphic is then translated to the target plane along the normal direction of the target plane, thereby improving the completeness and accuracy of the shape fitting.

Benefits of technology

It significantly improves the accuracy of grasping posture detection and the grasping success rate, reduces the impact of factors such as point cloud occlusion or noise, and ensures the accuracy and stability of grasping posture.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122125702A_ABST
    Figure CN122125702A_ABST
Patent Text Reader

Abstract

This application relates to a grasping posture detection method, system, device, robot, computer-readable storage medium, and computer program product. The method includes: extracting first position data and second position data from the position data of a target box; the first position data is located within a first distance threshold range of a first surface, and the distance between the first position data and the first surface is less than the distance between the second position data and the first surface; the first surface is perpendicular to the vertical coordinate axis in the physical coordinate system; performing shape fitting based on the first and second position data to determine a first contour pattern; performing principal component analysis on the first position data to determine a target plane, and translating the first contour pattern to the target plane along the normal direction of the target plane; and detecting the grasping posture of the target box based on the translated first contour pattern. This method can reduce contour fitting deviation, improve the accuracy of grasping posture estimation, and increase the success rate of grasping the target box.
Need to check novelty before this filing date? Find Prior Art