Strain gauges position based on machine vision positioning

Abstract

In this article, in order to meet the requirement of automatic positioning before strain gauge grinding, we proposed a new machine vision algorithm based on connected region extraction and minimum circumferential rotation rectangle delineation. At first, based on the location information of the strain gauge, we used algorithms for image binaryzation, and extracted the strain gauge block by using the connected component extraction method, next roughly positioned strain gauges. And then we dispersed the part which connected to the strain gauge, moreover, found the minimum external rotating rectangle of each strain gauge in the pair of strain gauges so as to located the center of the strain gauge and realize the strain gauge positioning in the end. Then the position of other strain gauge was inferred according to the position of strain gauge in the corner. Experimental verification showed that the error of the algorithm in this article is under 0.05mm, and the algorithm met the requirements of project positioning that the error has to below 0.1mm.