2D image based road pavement crack detection by calculating minimal paths and dynamic programming

Road distress needs to be detected early to optimize road maintenance cost; automatic survey of road distress is a big challenge, particularity for the detection of tiny cracks due to important variation of pavement textures. This paper presents a new method for crack detection by finding the minimal path passing on each pixel of image from every path with a length d; we propose also a dynamic programming implementation to make it applicable in real condition. Methods are tested on synthesis images set and a large set of real images. Results show that cracks as small as 2mm could be detected.