In order to resolve the problems of emitter identification caused by parameter complexity and agility of multi-function radars, S transform (ST) is introduced to the analysis of radar emitter signals and a new scheme for feature extraction of intra-pulse modulated radar signals is presented.Firstly the moment invariant features of the ST images are extracted.Secondly, aiming at the problems of low correct classification rate to BPSK signals and QPSK signals, a new approach to extract the features of peak envelope is presented. Finally, Support vector machines (SVM) was used to design classifiers for experimentation.The simulation results show that, the algorithm is of high and stable classification precision at the signal to noise ratio of wide range.%针对利用常规参数进行复杂体制雷达辐射源信号分类识别存在的问题,提出一种基于S变换的雷达脉内调制特征提取方法.将S变换引入雷达脉内信号分析中,首先提取信号S变换时频图像的不变矩特征;然后针对其在BPSK和QPSK信号分类上的不足,提出了基于S变换峰值包络特征提取方法;最后用支持向量机作为分类器进行分类实验.仿真结果表明,在一定信噪比范围内,提取的特征具有较好的聚集性和可分性,对6种雷达脉内调制信号取得了较好的分类正确率.
展开▼