An efficient crack detection method using percolation-based image processing

Crack detection on concrete surfaces is the most popular subject in the inspection of the concrete structures. The conventional method of crack detection is performed by experienced human inspectors by sketching the crack patterns manually. Some automated crack detection techniques utilizing image processing have been proposed. Although most of the image-based approaches pay attention to the accuracy of the crack detection results, the computation time is also important for practical use, because the size of the digital image reaches 10-mega pixels. In this paper, we introduce an efficient and high-speed method for crack detection employing percolation-based image processing. To reduce the computation time, we consult the ideas of the sequential similarity detection algorithm and active search (SSDA). According to the concept of SSDA, the percolation process is terminated by calculating the circularity midway through the processing. Moreover, percolation processing can be skipped for the next pixel depending on the circularity of neighboring pixels. The experimental result shows that the proposed approach is efficient in reducing the computation cost while preserving the accuracy of crack detection result.