AUTOMATIC AND EXACT CRACK EXTRACTION FROM CONCRETE SURFACES USING IMAGE PROCESSING TECHNIQUES

Abstract

  In this paper, we proposed a method for robust automatic crack extraction from noisy concrete surface images. The proposed method has two preprocessing steps for robust extraction. After two preprocessing steps, probabilistic relaxation is applied to overcome the problem of threshold selection and also to prevent noises, and improved locally adaptive thresholding is used to extract cracks exactly and automatically. It is possible to classify width of cracks extracted by our proposed method. The experimental results show that the proposed method is effective to extract cracks from actual noisy concrete surface images, compared to the conventional method and other thresholding techniques.