| 摘要: |
| 高频地波雷达由于观测环境复杂，存在监测范围广、船只目标机动等特殊性，使其在特定目标实时跟踪方面存在困难，而船只目标的持续跟踪正是地波雷达应该具有的功能，以便发挥所具有的全天候远距离早期预警的优势。本文对高频地波雷达船只目标航迹跟踪方法进行了综述，对目前的典型航迹跟踪方法进行了性能分析和比较，并提出了基于弧段检测的航迹长时跟踪新方法。该方法基于弧段数据建立船只运动模型的深度学习网络，并将船只目标的不同源特征结合地波雷达杂波特性进行融合决策，实现杂波背景下船只目标未来航迹的准确预测。该方法充分挖掘弧段信息，采用多源特征融合和递归校正机制，适用于在杂波环境且在航道附近存在众多干扰船只的情况下对机动目标航迹的实时稳定跟踪。为高频地波雷达在复杂干扰环境下特定目标的持续跟踪提供理论基础和方法指导，为充分发挥地波雷达在海上监视监测中的作用提供技术支撑。 |
| 关键词: 高频地波雷达，弧段检测，航迹跟踪，深度学习 |
| DOI： |
| 分类号:TN91 |
| 基金项目:国家自然科学基金(No.41506114；No. 61132005),海洋公益性科研专项项目(No.201505002) |
|
| Research on Specific Target Tracking Method Based on Arc Detection with HF Ground Wave Radar |
|
Zhang Ling,Liu Xu,Jiang Yi and Ji Yonggang
|
|
Ocean University of China,Ocean University of China,Ocean University of China,First Institute of Oceanography, State Oceanic Administration
|
| Abstract: |
| The ocean observation environment is complex with particular situations such as wide monitoring range, ship maneuvering and so on, which makes real-time tracking specific target becomes difficult for high frequency surface wave radar (HFSWR). The paper discusses the typical trajectory tracking methods and analyzes the performance of different methods. Furthermore, the vessel target tracking method based on arc detection of the trajectory is proposed. The new arc detection based trajectory tracking method is flexible and taps fully the arc segment information, which realizes the accurate extraction of the specific target trajectory even in the channel with a lot of interference ships. |
| Key words: HFSWR, arc detection, trajectory tracking, deep learning |