一种基于视觉算法的拆码垛定位方法及系统
By acquiring 3D scene data to generate scene version identifiers and candidate gripping surfaces, performing adsorption verification, and judging based on vacuum state, the problem of inaccurate vacuum adsorption positioning in existing technologies is solved, thereby improving the accuracy and reliability of robot depalletizing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 清研自动化技术(洛阳)有限公司
- Filing Date
- 2026-06-18
- Publication Date
- 2026-07-17
AI Technical Summary
Existing robot depalletizing and positioning solutions struggle to accurately determine the suitability of vacuum adsorption when exposed surfaces have local undulations, unclear boundaries, varying bonding gaps, or areas that do not meet airtightness requirements. This leads to positioning and execution failures, and the failure feedback cannot effectively correct the selection sequence.
By acquiring 3D scene data, a scene version identifier is generated, candidate grasping surfaces are formed and candidate identifiers are configured, and associated records are established. Based on the execution evaluation value and robot executable constraints, a test target is selected, adsorption verification is carried out, and the adsorption conditions are judged according to the vacuum state characteristics. If the conditions are not met, the execution evaluation value is reduced and the target is reselected or the data is reacquired.
This solves the problem of the inability to effectively connect candidate poses with physically adsorbable states, ensuring that adsorption verification results can be traced and that positioning selection can be corrected in a timely manner when the scene changes, thereby improving the accuracy and reliability of robot depalletizing.
Smart Images
Figure CN122401458A_ABST