Automated crack detection for concrete surface image using percolation model and edge information

A crack Is an important indicator in concrete structure diagnosis. Although cracks tend to display linear characteristics, it is not easy to detect them by conventional methods. There are a variety of difficult sources of noise: concrete bleb, stain noise, insufficient contrast, and shadings. In this paper, we introduce a novel crack detection method for a concrete surface image based on a percolation model. Our method evaluates the central pixel in a local window according to a cluster formed using percolation processing. In addition, we describe reducing the computational burden while still preserving the precision of the crack detection. To achieve this, our method utilizes edge information to reduce the number of starting points for percolation processing. The validity of the proposed technique is investigated through experiments with images of real concrete surfaces and it is shown that robust and reliable crack detection without oversight is achieved.