Image-enhanced YOLOv5 and Deep Sort Underwater Multi-moving Target Tracking Method

Underwater target tracking has a wide range of applications in the fields of fishery resources research, marine ecological environment exploration and AUV/ROV operating in the deep water. The target tracking method based on deep learning is applied to the underwater environment, the YOLOv5 detection algorithm is used as the detector to identify moving fish under the water and the Deep Sort tracking algorithm is combined to realize the target detection and tracking task. According to the characteristics of complex underwater light conditions and low target contrast, the collected underwater images are enhanced. The experimental results show that the tracking method has a higher accuracy in detecting fish targets, a better tracking effect and a higher real-time tracking performance than YOLOv3, it's dection accuracy can reach 96 %.