Automatic Pavement Crack Detection Algorithm

Abstract

Pavement crack detection and evaluation is a vital component in highway maintenance and management. It is
difficult to detect cracks effectively and accurately through ideal detecting algorithm. This paper improves the pavement
crack image processing algorithm from the angle of grey level transformation, median filter and image intensification
according to the characteristics of the pavement crack image. Utilizing this model, this study implements an automatic crack
classification and quantification method for defect image segmentation and image edge detection, results show that the
proposed detection method can effectively remove the isolated noise point, smooth the edge and improve the segmentation
accuracy.