基于时频分布图像和主分量分析的脉内调制识别算法研究

Study on intra-pulse modulation recognition method based on time-frequency distribution image and principal component analysis

本文提出了一种利用模式识别中图像处理的方法来进行雷达脉内调制特征识别的新算法。通过对雷达信号时频分布图像进行二维小波分解,并对其进行主分量分析,获得每类雷达信号中的特征参数,构建相应的分类器,并利用此分类器对信号进行识别。仿真结果表明,利用信号的时频分布图像来进行信号识别是完全可行的,提取的特征参数具有很好的鲁棒性,具有较强的抗噪声能力,同时可以取得较高识别率。

This paper proposes a new method for radar intra-pulse modulation characteristic recognition by means of image processing method in the field of pattern recognition. Through two-dimensional wavelet decomposition for time-frequency distribution image and principal component analysis, the feature parameters of each type radar signal can be obtained, and constructed the corresponding classifiers, which can be used for signals identification. The experiment results show it is feasible completely for signals recognition using time-frequency image of signals, and the extracted feature parameters have good performance, and also have higher recognition rate.