Surface crack detection using image processing and support vector machine

Surface crack detection has proven crucial in the process of inspecting civil engineering structures, as well as in many manufacturing industries that require high-quality products. A large number of methods that deal with this problem are based on computer vision, due to its effectiveness and affordability. In this paper, we present an end-to-end solution that uses the Canny edge detector combined with the Hough transform for feature extraction. With an aim to keep the process intuitive and have fast training and inference, we used SVM (support vector machine) for classification.