As a large oceanic country, our country has abundant marine resources. The detection of marine objects plays an increasingly important role in marine economic activities. Due to the frequent incidents of sea area disputes in recent years, it is of far-reaching practical significance to improve my country's sea surface monitoring technology near the sea. Maritime target detection and tracking technology will play a huge role in the field of sea surface monitoring. This paper will take ships at sea as the main research object, and study the method of multi-target monitoring and tracking at sea based on deep learning. By improving the YOLOv5 algorithm, integrating multiple attention mechanisms, and matching the DeepSORT tracking algorithm, real-time detection and tracking of maritime targets is realized. Not only that but to achieve better detection results, this paper manually created a data set of maritime targets. Through the collection of data images such as ships, buoys, islands, and reefs, a rich data set of maritime targets has been established, which can be used for better training with algorithms.
Multi-target tracking and detection on sea surface based on YOLOv5
2024-05-17
1992579 byte
Conference paper
Electronic Resource
English
Multi-target Detection in Airport Scene Based on Yolov5
IEEE | 2021
|Multi-scale Small Target Detection for Indoor Mobile Rescue Vehicles Based on Improved YOLOv5
Springer Verlag | 2024
|Multi-scale Small Target Detection for Indoor Mobile Rescue Vehicles Based on Improved YOLOv5
Springer Verlag | 2024
|STD-Yolov5: a ship-type detection model based on improved Yolov5
Taylor & Francis Verlag | 2024
|