关键词: 脉内调制, 支持向量机, 特征提取, 时频图像

Abstract: In order to achieve the automatic recognition of the pulse modulation of radar, a new method to simplify the figure of the results of time-frequency analysis to extract features was proposed, and then Support Vector Machine (SVM) was used to sort them according to the features. Firstly, extracted the points with useful information; secondly, found the center of each column of useful points; lastly, sampled these centers at the same length, so the figures were transformed to curves with the same length, and the dimensions of the feature of SVMs had been sharply decreased. The simulation results show that the pulse modulation of radar can be recognized with a high accuracy by the algorithm of this paper. Additionally, this method can tolerate the noise, and the accuracy can be kept at a high level while the SNR (Signal-to-Noise Ratio) is low.

Key words: pulse modulation, Support Vector Machine (SVM), feature extraction, time-frequency figure