Improved YOLO-based algorithm for urban traffic object detection

Urban traffic vehicle detection is a key component of smart city transportation systems, aimed at improving traffic management and safety through modern technologies and information methods. In response to the characteristics and challenges of vehicle detection in smart cities, this paper proposes a vehicle detection method based on drone aerial images, employing an object detection algorithm based on YOLOv4, namely the Adaptive Cropping YOLO algorithm. Through training and optimization on a large-scale dataset, this method can accurately detect and identify different types of urban vehicles. Experimental results show that this algorithm can effectively detect large-sized image targets that traditional YOLO algorithms may miss, providing reliable technical support for traffic safety monitoring and management.