A new crack detection method for bituminous pavements based on cluster and minimum spanning tree is proposed in this paper. Firstly

the Retinex model was used to dodging preprocessing for crack images

to reduce the difference of different lighting conditions

and then the edge detection and image segmentation technology was both used for extracting the crack area in the image

finally the clustering and minimum spanning tree technology was used for crack detection. The test images used in this study are captured from actual traffic intersections. The experimental results show that the method can accurately and effectively detect the crack for bituminous pavements

and can be applied on real traffic scene with considerable practical value.