Positioning control system based on computer vision

摘要

This article analyzes the actual case in practical engineering field of transferring goods from an irregular tray where objects are placed in each section within a relatively large tolerance to a fixing fixture through a robot. Based on the transformation of a Cartesian coordinate system, it can compute the optimized movement of the robot in order to complete the needed rotation and translation in one single operation. Computer vision system and robot motion control are combined. The two-dimensional Cartesian coordinate systems of computer vision and robot need to be matched in order to the let the two systems work together. At the same time calibration and matching of the measurement units of the two systems as well as the compensation of image distortion play a key role in the practical implementation. This solution leads to a precise and time-efficient positioning.