Abstract: With regard to the problem of declining of road surface image quality in high-speed processing, the grey entropy is applied to the road surface image enhancement to lay a good foundation for the automatic detection of cracks. We use the grey entropy of the texture of the local area in the road surface image to characterize the scale of the increase or decrease in the process of image local contrast enhancement and fine-tune the local contrast function for that we could effectively maintain the smooth region relatively smooth while enhancing the contrast of the image edge region. Finally, simulation results show that the algorithm is more effective to improve the quality of pavement crack image than other traditional algorithms and it is a good algorithm.