A Vision-based High-precision Positioning and Error Compensation Method for Adsorptive Embedded Components

To address the alignment deviation between the suction cup and embedded component centers during absorptive assembly of wrapped embedded components, a vision-based positioning and error compensation method integrating edge template matching with multi-circle marker geometric constraints is proposed. A template matching algorithm based on edge gradient direction consistency is developed, combining an image pyramid hierarchy with a genetic algorithm for decoupled parameter search to achieve robust end-face center localization under complex conditions. For the occluded suction cup center, a concentric double-ring multi-circle marker is designed, and a two-step single-circle fitting method with multi-circle joint optimization is developed, together with inverse projection transformation to correct projection deviation. Experimental results demonstrate that template matching time is reduced to 0.63 s and positioning error is controlled within 0.01 mm, with robust performance maintained under varying illumination and partial occlusion, providing reliable error compensation for high-precision absorptive assembly embedded components.