Machine Visual Location Method Research in Robot Unstacking

Abstract

Work pieces’ location based on machine visual in robot unstacking job has been researched. Adaptive threshold segmentation and modified Hough Transform are used to extract the target region needs further processing. A new method is proposed to choose the appropriate threshold to binary different images with different brightness. Watershed algorithm and Morphology algorithm are used to detect and connect the edges of work pieces. With the distance sensor and camera calibration to complete the location of pallet and work pieces, guiding robot to finish unstacking job successfully.