Real-Time Underwater Maritime Object Detection in Side-Scan Sonar Images Based on Transformer-YOLOv5

To overcome the shortcomings of traditional manual detection underwater targets in side-scan sonar (SSS) images, a real-time automatic target recognition (ATR) method is proposed this paper. This consists image preprocessing, sampling, ATR by integration transformer module and YOLOv5s (that is, TR–YOLOv5s), localization. By considering target-sparse feature-barren characteristics SSS novel TR–YOLOv5s network down-sampling principle are put forward, attention mechanism introduced to meet requirements accuracy efficiency for recognition. Experiments verified achieved 85.6% mean average precision (mAP) 87.8% macro-F2 score, brought 12.5% 10.6% gains compared with trained from scratch, had speed about 0.068 s per image.