Research on Neural Network Algorithm for Road Crack Detection

Abstract

Aiming at the problem of road crack detection, experiments based on CFD data set and Unet model were carried out. The overall structure of the network and the contribution of each module to the final detection effect of Unet model were analyzed. Some modules in Unet model were cut and replaced based on that, and CCUnet model was obtained. Final experiment results show that detection rate of CCUnet model is greatly improved under the condition that model accuracy is maintained basically, and that has practical application value.