Crack Detection Algorithm of Complex Bridge Based on Image Process

Abstract

With highway bridge construction developing rapidly, bridge safety is crucial. Crack detection efficiency directly affects bridge safety and service life. Therefore, digital image processing for crack detection technology is important. Existing crack detection algorithms only detect single, regular cracks but not multiple cracks in complex, extended directions. This paper proposes detecting cracks in images by adopting OTSU automatic threshold, guided filtering, and gamma image enhancement, then using Zhang Suen skeleton extraction algorithm to extract crack skeletons. Hough line detection is conducted to detect different trends of multiple cracks, and scanning line algorithms is used to calculate normal crack width with engineering significance. The result of the research shows that the algorithm detects cracks efficiently, and the thinning algorithm based on skeleton extraction extracts crack trends accurately. The scan line algorithm has practical significance for width measurement of irregular cracks. Therefore, this algorithm has strong application value in bridge crack detection.