| 摘要: |
| 针对窄带雷达旋翼微动特征参数提取问题，研究了一种基于闪烁现象的旋翼微动特征提取方法。首先通过自相关处理估计出旋翼目标回波的闪烁周期，并利用闪烁周期与叶片数的关系估计得到多种可能的转速，其次通过时频分析及骨架提取方法，判断叶片的奇偶数及估计旋翼目标的最大微多普勒频率，并利用先验区间对估计值进行初步筛选，最后在保持与原始信号中噪声功率相同情况下，利用估计值对回波信号进行重构，从能量差异的角度对估计值进行第二次筛选，将原始信号能量与重构信号能量最接近的估计值输出作为最终的估计值。仿真结果验证了本文方法的可行性和有效性。 |
| 关键词: 微动 闪烁 自相关 骨架提取 特征提取 |
| DOI：DOI:10.3969/j.issn.1672-2337.2019.06.014 |
| 分类号:TN958.3 |
| 基金项目:国家自然科学基金(No.61671469) |
|
| Feature Extraction Algorithm for Micro Motion Target Based on Flash Phenomenon |
|
XIA Saiqiang，ZHU Mingshuo，CHEN Wenfeng，YANG Jun
|
|
Air Force Early Warning Academy，Wuhan 430019，China
|
| Abstract: |
| Aiming at the extraction problem of micro motion parameters in narrow band radar，an extraction algorithm for micro motion target parameters based on flash phenomenon is studied. Firstly，the flash period of the rotor target is estimated by autocorrelation processing，and multiple possible rotation speeds are estimated by the relationship between the flash period and the number of blades. Secondly，the odevity of the blade and the maximum micro Doppler frequency of the rotor target are estimated by time frequency analysis and skeleton extraction method. Then，a priori interval is used to screen the estimated values preliminarily.Finally，the signal is reconstructed by using the estimated value in the case of the same noise power with the original signal，and the estimated value is secondly filtered from the perspective of the energy difference.The estimated value，obtained when the energy of the reconstructed signal is most close to the energy of the original signal，is taken as the final estimation.The simulation results verify the feasibility and effectiveness of the proposed algorithm. |
| Key words: micro motion flash autocorrelation skeleton extraction feature extraction |
