With the highway bridge construction develop rapidly,as an important link of traffic hub, the safety of bridge has always been the focus of people's attention.As an important component of the bridge health monitoring, the service life of bridges would be affected directly by the efficiency of crack detection. Therefore, the crack detection technology based on digital image process has become a hot issue in research as well as the difficulty. Existing algorithms of crack detection could only detect the single and regular cracks but do not satisfy requirements of extracting and detecting multiple cracks in the case of complex extension direction. In order to solve this problem, this paper would detect the cracks in the image by adopting OTSU automatic threshold, guided filtering, gamma image enhancement and other methods, and then use Zhang_Suen skeleton extraction algorithm to extract the crack skeleton, the method of hough line detection would been conducted to detect the different trend of multiple cracks, and finally use the scanning line algorithm to calculate the normal crack width with engineering significance. The research result of the shows that cracks in images could be detected efficiently by the algorithm, and the thinning algorithm based on skeleton extraction could extract the trend of cracks accurately. The scan line algorithm has practical significance for the width measurement of irregular cracks. Therefore, this algorithm has a strong application value in the detection of bridge cracks. Keywords:Crack Detection,Skeleton Extraction Algorithm,Scan Line Algorithm, Normal Crack Width.

CICTP