Morphology and maximum entropy image segmentation based urban pavement cracks detection

In order to solve the problem of big workload in the post-processing and more image noises in the urban pavement image, using morphology and the maximum entropy image segmentation, a method to detect urban pavement crack is proposed. After the initial classification, based on morphology with different scales and different directions of elements, this study uses opening-by-reconstruction to enhance gray difference between the crack target and the pavement background, then uses the maximum entropy image segmentation to make image segmentation, and further process the binary image, and lastly classifies crack according to the principle of projection analysis. The experimental results show that the synthetic method obtains a better detection result, rapidly and accurately detecting the edge of the crack while effectively suppressing the noise.