基于CWT的常规雷达编队目标架次自动识别

Formation Radar Targets Automatic Detection Method Based on Choi-Williams Trispectrum

基于多普勒效应的编队目标架次识别的关键,在于要有噪声抑制和频率分辨能力良好的特征提取算法。文中把在乘性和加性高斯噪声背景中具有优良的提取调频信号分量能力的Choi-Williams三谱(CWT)引入编队目标架次识别特征提取,构建了基于CWT三谱的距离-时间-多普勒投影像形态特征的架次自动识别方法。对实测编队目标回波CWT像形态特征的架次识别实验说明,该特征提取算法具有良好的噪声抑制能力和频率分辨率;信噪比为-5 dB的条件下对二架和四架编队目标的正确识别率不低于97%。

Favorable noise suppressing and frequency resolving power is the key of Doppler based formation radar targets detection. The Choi-Williams Trispectrum （ CWT）, defined as a member of time-varying higher-order spectra based on the polynomial Choi-Williams distributions, is a powerful time-frequency method to analyze FM signals in multiplicative and additive gaussian noise with adequate frequency resolving power. A formation radar targets detection method based on the morphologic features of CWT range- time-Doppler projection profile is proposed. Tests of field conventional coherent radar echoes verify the validity of the proposed method. Two or four targets are detected separately with an accuracy not less than 97%, while SNR = -5dB.