Research on Crack Edge Detection of Aircraft Skin Based on Traditional Inspired Network

Aircraft skin crack is a common type of aircraft skin damage, and the traditional detection method in aircraft skin crack effect is not good. Along with the deep learning is widely used in areas such as target detection, based on Deep Neural Network, we introduce a traditional method inspired framework to produce good edge with minimum complexity. We simplify the network structure by adding Feature Extractor, Enrichment and Summarizer, which are basically consistent with the gradient, low-pass filtering and pixel connection in the traditional edge detection scheme. The structure can effectively reduce the complexity of edge prediction and maintain the quality of edge prediction.