A Robust Method for Automatically Detecting Cracks on Noisy Concrete Surfaces

Automatic crack detection from concrete surface images is very effective for nondestructive testing. In our previous work, we proposed two preprocessing methods for automatic crack detection from noisy concrete surfaces. In this paper, we propose an automatic crack detection method after the preprocessings. The proposed method consists of two steps. One is relaxation process to prevent noises, and the other is a improved locally adaptive thresholding to detect cracks exactly. We evaluate the performance of the proposed method using 60 actual noisy concrete surface images, compared to those of the conventional thresholding techniques. Experimental results show that the robustness and the accuracy of the proposed method are good.