Binocular measurement method for the continuous casting slab model based on the improved BRISK algorithm.

Abstract

Due to the low accuracy of the traditional image feature matching algorithm in binocular vision measurement, a binocular measurement method for the continuous casting slab model based on the improved binary robust invariant scalable keypoints (BRISK) algorithm is proposed. First, the feature points of the image are detected. After that, local area sampling and sub-area division are carried out with the feature points as the center, sub-areas with low offset values are removed, and the main direction is obtained by using the centroid of the remaining sub-areas. Then, the gray difference threshold is used to replace the traditional gray value intensity comparison to generate descriptors. Finally, the Hamming distance is used to match the feature points, and the three-dimensional coordinates of the matching points are calculated to complete the measurement. Through comparative experiments, the lowest relative error of the improved algorithm in this paper reaches 0.4723%, which meets the requirement of measurement accuracy.