Research on ship target detection based on improved YOLOv5 algorithm
Anni Zhang, Xia Zhu
Abstract
Aiming at the problem that the YOLOv5 algorithm has low detection accuracy for small targets and is prone to missed detection and false detection, this paper proposes an improved YOLOv5 algorithm by introducing the coordinate attention mechanism and the bidirectional feature pyramid network for effective waterborne ship detection. This method aims to improve the detection accuracy of waterborne targets. In this study, we focus on constructing a dataset of waterborne ship ships, modifying the YOLOv5 algorithm, and using the PyTorch framework for experiments to evaluate the performance of the proposed method. The experimental results show that the average accuracy of the improved detection algorithm is 99.1%, which is 3.3% higher than that of the original YOLO v5.