Soccer Player Tracking Using UAV Imagery: A Comparative Study of Yolo and Traditional Image Processing Algorithms

Abstract

Player tracking is a useful tool for tactical analysis and performance evaluation in soccer, providing valuable insights into player movements and team dynamics. This project investigates the feasibility of tracking players using UAVcaptured imagery, employing both YOLO and traditional image processing algorithms (TIPA). Initial validation focuses on robot soccer players due to their predictable and controllable movements. The comparative analysis considers processing time, computational cost, adaptability to environmental changes, sensitivity to lighting variations, ability to handle dynamic conditions, tracking accuracy, and real-time performance. Results indicate that, under equivalent hardware and preparation time conditions, YOLO achieves performance comparable to traditional techniques. Nonetheless, the selection of the most suitable approach should be guided by task-specific demands, available computational resources, and the time allocated for system development and deployment.