A supervised approach using transformer networks for the detection of turning-related anomalies in urban intersections

Traffic anomalies in urban areas are the main reason for traffic congestion and accidents. an effective tool for decreasing these anomalies is employing intelligent systems for monitoring road traffic and law enforcement. these systems may make decisions using features based on different modalities. to detect anomalies, video-based systems usually use the speed, direction, and trajectory of vehicles. among various types of traffic anomalies, this research focused on turning-related anomalies in urban roads and intersections, which are the most critical challenges of intelligent traffic monitoring systems. in this research, a supervised system based on a transformer network was designed and examined on real traffic video sequences. the evaluation of our system demonstrated that the proposed method can detect turning-related anomalies with about 96% accuracy.