Study on Expressway Crack Segmentation Algorithm Combined with Pulse-Coupled Neural Network and Cross-Entropy Algorithm

摘要

The expressway crack identification is significantly important for the expressway safety maintenance, and the crack detection is one of key technologies for the crack identification. This paper proposed an expressway crack detection method based on improved Pulse-Coupled Neural Network (PCNN), used minimum cross-entropy algorithm to obtain the optimal iterations of PCNN algorithm, and then complete the segmentation of expressway images by combining the simplified PCNN algorithm. The results showed that this method could inhibit the background noise and better extract continuous crack edge to provide good characteristics for crack identification in the next step.