Radar Emitter Sorting and Recognition Based on Time-frequency Image Union Feature

This paper proposed a radar emitter sorting and recognition algorithm based on Choi-Williams Distribution (CWD) time-frequency image union features for problem of low recognition rate and high complexity of radar signals soring and recognition algorithm at low signal-to-noise ratio (SNR). Firstly, the time-frequency image matrix of signals is obtained via CWD, then the time-frequency image is processed to extract the Gray Level Co-occurrence Matrix (GLCM), Zernike moment features and singular value entropy features of time-frequency image to form union features parameter vectors, and finally the automatic sorting and recognition of radar signals is achieved via support vector machine (SVM) classifier. Simulation results show that the overall average recognition rate of the proposed method for 6 types radar signals could reach above 97.3% under -6dB SNR, it also has a better recognition effect on the hybrid modulated signal.