Underwater Target Detection Algorithm Based on Improved YOLOv5

Underwater target detection plays an important role in ocean exploration, to which the improvement of relevant technology is much practical significance. Although existing algorithms have achieved excellent performance on land, they often fail achieve satisfactory outcome when underwater environment. In this paper, one most advanced algorithms, YOLOv5 (You Only Look Once), was first applied environment before being improved by combining it with some methods characteristic To be specific, Swin Transformer treated as basic backbone network YOLOv5, makes suitable for those images blurred targets. It possible focus fusing relatively resolution features improving method path aggregation (PANet) multi-scale feature fusion. The confidence loss function basis different layers, biased learn high-quality positive anchor boxes and make more capable detecting target. As suggested experimental results, model effective targets, mean average precision (mAP) reaching 87.2%, advantageous over general models fit use complex