A robotic arm active visual optimization grasping method and system for a complex occlusion environment
By constructing a 3D voxel map and optimizing the viewpoint using the information entropy gain algorithm, and combining it with the ICP algorithm for point cloud registration, the problem of pose estimation accuracy and efficiency of robotic arm visual grasping under complex occlusion environments was solved, achieving a high success rate of grasping.
CN122231884APending Publication Date: 2026-06-19JINCHENG ZHIXING (CHENGDU) INTELLIGENT TECHNOLOGY CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JINCHENG ZHIXING (CHENGDU) INTELLIGENT TECHNOLOGY CO LTD
- Filing Date
- 2026-04-28
- Publication Date
- 2026-06-19
Smart Images
Figure CN122231884A_ABST
Abstract
This invention relates to a method and system for active vision-based optimal grasping of a robotic arm in complex occluded environments, belonging to the field of industrial robot technology. The method includes: Step S1, acquiring point cloud and image data of the work scene; Step S2, calculating the confidence score of the current pose estimation, and proceeding to Step S3 when the confidence score is lower than a threshold; Step S3, generating an optimal viewpoint command; Step S4, extracting the end-effector pose parameters from the optimal viewpoint command to complete the conversion from the abstract viewpoint command to a joint space trajectory sequence; Step S5, point cloud fusion and feedback. This invention solves the problems of inaccurate pose estimation, low observation efficiency, and large fluctuations in grasping success rate caused by occlusion in existing technologies, achieving high-success-rate grasping in complex occluded environments.
Need to check novelty before this filing date? Find Prior Art