Real-time Underwater Target Tracking Using PP-YOLO and Cloud Computing

Abstract

With the growing interest in ocean exploration, accurately tracking underwater targets has become increasingly important for resource exploitation and environmental protection. This paper explores the application of deep learning algorithms for multi-target tracking in underwater environments. The challenges of image processing in this context are discussed, and the YOLOv3 target detection algorithm is utilized to train a real-time underwater target tracking model with image enhancement techniques. Furthermore, the advantages and disadvantages of the YOLOv3 and PP-YOLO algorithms are compared by training the PP-YOLO model in the cloud. This study contributes to the development of more efficient and reliable methods for underwater target tracking.