Road Disease Detection Algorithm Based on Probability Overlay Edge Detection Network

• 1、School of Cyberspace Security, Being University of Post and Teleommunication, Beijing, 100876

Abstract：The detection of road diseases plays an important role in the maintenance of roadways and forms the basis for promoting healthy and regular progress in the construction of transport infrastructure. However, the traditional method of manual detection has been difficult to meet the current real-time detection requirements of the huge road network. Therefore, research and design of an efficient, precise and robust method for detecting road diseases is of great importance for the maintenance of pavements. This article focuses on the automatic detection of roadway diseases on the basis of digital images, and proposes a road disease recognition algorithm based on a probabilistic network for detection of overlapping edges, which increases the probability gain generated. by the complex features of the disease based on the original probability map. Experimental results show that the recognition accuracy of this algorithm is much higher than that of traditional methods, especially for the recognition of particular diseases.

Keywords： crack detection convolutional neural network two-channel