Rotating Target Detection for Tarpaulin Rope Based on Improved YOLOv5

Waist rope is the main restraint on the tarpaulin when the train is running. Aiming at the problem of low accuracy for waist rope detection by the traditional horizontal frame detection approaches, this paper studies the principle of rotating target detection, proposes an effective detection method for the waist rope based on the improved yolov5 deep learning algorithm by computer vision technology and Gaussian window function, so as to realize the detection of tarpaulin rope with different angles. Experiments show that the proposed method has high detection accuracy for waist ropes and can accurately reflect the position information of tarpaulin waist ropes.