This paper proposes a technique for detecting and analyzing cracks in the surface of concrete. Based on digital image processing, the technique automates the measurement of features such as the width, length, and direction of cracks. The proposed algorithm has a new way of determining the unit pixel length, it applies a morphological technique for shading corrections, it uses enhanced binarization and shape analysis for better detection, and it suggest algorithms for calculating the width, length, and direction of cracks. A MATLAB code was developed for the proposed algorithm and, to determine the effectiveness of the algorithm, testing was performed on crack images taken with a digital camera. Within limited testing, the accuracy of the detection and analysis of cracks is better with the proposed method than with conventional methods. This paper presents a preliminary study for commercial development.
展开▼