| 引用本文: | 袁 雪， 韦楠楠， 张兴敢. 利用低数据率HRRP序列进行弹道中段目标识别[J]. 雷达科学与技术, 2023, 21(5): 559-567.[点击复制] |
| YUAN Xue, WEI Nannan, ZHANG Xinggan. Ballistic Midcourse Target Recognition Based on Low Data Rate HRRP Sequence[J]. Radar Science and Technology, 2023, 21(5): 559-567.[点击复制] |
|
| 摘要: |
| 雷达目标识别是弹道防御阶段的关键环节，为节约雷达时间资源和降低对计算机处理能力的要求，需研究低数据率雷达回波信号的弹道中段目标识别方法，本文以低数据率目标高分辨一维距离像序列（High Resolution Range Profile，HRRP）为研究对象，提出了基于图像投影法的进动频率特征提取算法和基于特征级融合的弹道中段目标识别方法，解决了由于HRRP回波序列数据率过低而导致的时频曲线周期模糊和单一特征造成目标识别准确率浮动大的问题。本文通过仿真弹道导弹中段飞行场景中弹头、重诱饵、轻诱饵、碎片目标的特性数据，同时考虑目标尺寸、形状和微运动模型等差异，结合仿真数据对本文所提算法进行验证。实验结果表明在低数据率（10~100 Hz）下，HRRP序列利用本文算法提取的进动频率特征结果误差值小于0.05 Hz，具有较高准确性和稳定性，通过特征融合方法联合进动频率和目标结构特征将弹道中段目标的识别准确率提升到了96%以上且趋于稳定。 |
| 关键词: 弹道中段目标识别 低数据率一维距离像序列 进动特征提取 特征级融合 |
| DOI：DOI:10.3969/j.issn.1672-2337.2023.05.013 |
| 分类号:TN957 |
| 基金项目: |
|
| Ballistic Midcourse Target Recognition Based on Low Data Rate HRRP Sequence |
|
YUAN Xue, WEI Nannan, ZHANG Xinggan
|
|
School of Electronic Science and Engineering, Nanjing University, Nanjing 210023, China
|
| Abstract: |
| Radar target recognition is a critical step in ballistic missile defense. In order to save radar time resources and reduce the requirements on computer processing power, it is necessary to study the ballistic midcourse target identification method of radar echo signal with low data rate. This paper takes the high resolution range profile（HRRP） sequence with low data rate as the research object, and proposes a precession frequency extraction algorithm based on image projection method and a ballistic midcourse target recognition method based on feature?level fusion. The methods solve the problem of large fluctuation of target recognition accuracy due to the fuzzy period of time?frequency curve and single feature caused by low data rate of HRRP echo sequence. The experiment firstly simulated the characteristic data of warhead, heavy decoy, light decoy and fragment target in the midcourse flight scenario of ballistic missile. Meanwhile, considering the differences of target size, shape and micro?motion model, the algorithm was analyzed by combining the simulation data. Experimental results show that at low data rate （10~100 Hz）, the error value of precession frequency feature extracted by the algorithm in this paper is less than 0.05 Hz, which has high accuracy and stability. By combining precession frequency and target structure characteristics with feature fusion method, the recognition accuracy of ballistic midcourse target is improved to more than 96% and tends to be stable. |
| Key words: ballistic midcourse target recognition low data rate of HRRP precession feature extraction feature⁃level fusion |
