Abstract:

Aiming at the problem of ingot stack orientation on high-speed continuous casting aluminum ingot production line,we propose an aluminum stack positioning method,which obtains pose information using binocular vision.We take the steel belt of the aluminum ingot stack as the feature information,use the SURF feature extraction algorithm to identify the target of the same feature point.We optimize the feature points and perform distance constraint on them.We use the corner coordinate points obtained through feature matching and the binocular ranging formula jointly to calculate the pose information of aluminum stacks.

Key words: binocular vision, feature extraction, image matching, SURF