A Shallow Crack Detection Method Based on Attractor Model

When the overall image processing method is used to detect the cracks in the building with different degrees of depth, the shallow cracks could not be detected or the detection is incomplete. In this paper, a shallow crack detection algorithm based on attractor model is proposed. Each pixel on the image is assumed to be a vector and regarded as an attractor. Then according to the degree of attraction between them, the two are matched to obtain multiple attractors. The rectangular region between each set of attractors was determined, and the corresponding rectangular region in the original drawing was processed, the final shallow crack skeleton diagram was obtained by splicing. The results show that this method can detect shallow cracks clearly and accurately.