Rotary workpiece docking method and device based on multi-view vision and pose optimization
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING INST OF TECH
- Filing Date
- 2025-09-29
- Publication Date
- 2026-07-21
AI Technical Summary
Traditional vision-guided methods, especially under large positional deviations, suffer from insufficient axis alignment accuracy in the automated assembly of rotating workpieces, making it difficult to accurately calculate the six-degree-of-freedom positional parameters.
A method based on multi-view vision and pose optimization is adopted. Workpiece information is acquired through a binocular stereo camera array, multi-source data fusion is performed, the cylindrical surface and end face features of pins and pin holes are extracted, the axis parameter equation is established, and the initial pose is iteratively calculated by combining a six-degree-of-freedom pose optimization model and a kinematic model. Closed-loop correction is performed through visual and encoder feedback information, and finally a force control sensor is introduced for compliant control.
It achieves high-precision docking of workpieces under large positional deviations, improves axis alignment accuracy, solves the robustness and accuracy problems of traditional methods under large deviation scenarios, and ensures stable docking of the robot arm.
Smart Images

Figure CN121132657B_ABST
Abstract
Citation Information
Patent Citations
Method and system for using vision to correct errors of metallurgical plug-in assembly robot
CN111571596A
Intelligent garbage sorting system and method based on visual grabbing of mechanical arm
CN120306273A