中文摘要：

近年来对压缩感知理论的研究,进一步证明了信号的稀疏表示方法在信号压缩、特征提取等方面的有效性及巨大的应用潜力。作为信号处理领域的典型应用之一,雷达目标识别已有许多成熟的算法,其中一些基于高分辨距离像进行识别,但是这些方法大多忽略了高分辨距离像信号自身的稀疏特点。为此提出了一种基于压缩感知稀疏分解实现高分辨一维距离像目标识别的算法。此算法首先构建组合正交冗余字典,在满足信号表示准确性的情况下,兼有正交字典运算快捷的特点;然后,通过改进的分组匹配稀疏分解算法,根据距离像训练样本快捷地求取其类别字典;最后,基于类别字典对测试样本进行分类实现目标识别。仿真实验证明该目标识别算法简捷、识别率较高、抗噪能力强。

英文摘要：

In recent years,with the development of compressed sensing theory,sparse representation is widely used in signal compression and feature extraction. This method presents tremendous application potential. Radar target recognition is one of the classic applications of signal processing and there are many recognition algorithms. Some recognition algorithms are based on high resolution range profile（ HRRP）,but less of them employ the sparseness of HRRP samples. Thus,a radar HRRP target recognition algorithm based on sparse decomposition in compressed sensing is presented here. First,several orthogonal bases are used to compose a redundant dictionary which can satisfy the accuracy and speediness of HRRP sparse representation. Then,the training samples＇ taxonomic dictionaries are acquired by an improve grouping MP decomposition algorithm. Finally,the reconstruction errors of testing samples were calculated to recognize the targets. The simulation results show that this algorithm has higher recognition rate and better denoising performance. It is easy and practical for radar target recognition.