An inland waterway traffic complexity evaluation method using radar sequential images

Abstract

This paper proposes an inland waterway traffic complexity evaluation method using radar sequential images, which can be used to assist traffic situation awareness by maritime supervisors and reduce their workload. The proposed method consists of four steps: ship detection, ship tracking, ship interaction determination, and traffic complexity evaluation. First, target ships are detected through background subtraction to suppress the interference of radar noise in radar images. Second, the Kalman filter and Hungarian algorithms are introduced to match ship positions in two consecutive sampling intervals. Third, the interaction between ship pairs is determined by integrating the relative distance, encounter angle, and encounter trend. Fourth, ship pairs are marked when the interaction exceeds a set threshold, and a traffic graph is constructed to evaluate traffic complexity using the Edge Density and Strength indicators. The proposed method is tested on selected inland waters of the Yangtze River. The experimental results indicate that the proposed approach can evaluate traffic complexity through radar sequential images in real-time. The proposed approach is interactive and interpretable for maritime supervisors, and it helps them to quickly focus on particular waterways with potential risks.