摘要
因目标所处环境及探测用宽带雷达工作机理等因素的制约,实测距离像样本不仅繁多且可能混有干扰,为克服因此带来的识别困难,优化识别效果,稀疏分析提供了一类压缩样本数据并提升数据分析质量的研究思路。提出了一种依据分组稀疏分析策略开展雷达一维距离像稀疏分解,从而识别目标的方法。为了进一步改善识别算法的噪声鲁棒性,提升其实用性,算法在样本稀疏分析环节依信噪比开展了分解参数的调整。实验结果表明:对比同类识别算法,文中方法求解过程简洁,适用范围也相对更广;而较之一些不同类型的常规算法,文中方法则具备更好的噪声鲁棒性及更高的识别率。
Because of the complexity of environment factors and the working mechanisms of wideband detection radar,there are a tremendous number of real high resolution range profiles(HRRP)samples which is disturbed. To overcome the resulting identification difficulties and optimize the recognition result,sparse representation is an effective way to compress HRRP samples and improve the effect of samples analysis. In this paper,a partition sparse decomposition algorithm are introduced to implement the target identification of HRRP. In order to have better noise robustness and improve the practicability of the algorithm,a sparse decomposition parameter is adjusted on SNR during the sample sparse decomposition. The simulation results show that compared with similar recognition algorithms,the algorithm is more concise and applicable. Compared with some different conventional recognition method,it has better noise robustness and higher recognition ratio.
作者
段沛沛
雒明世
DUAN Pei-pei;LUO Ming-shi(School of Computer Science,Xi’an Shiyou University,Xi’an 710065,China)
出处
《火力与指挥控制》
CSCD
北大核心
2022年第1期15-19,共5页
Fire Control & Command Control
基金
国家自然科学基金(61171155)
陕西省科技厅一般基金资助项目(2018SF-409)。
关键词
雷达自动目标识别
高分辨距离像
稀疏分析
正交过完备组合字典
radar target recognition
high resolution range profile
sparse analysis
orthogonal redundant combinatorial dictionary
