多级奇异值分解和SG的通信雷达信号降噪方法

Noise Reduction Method of Communication Radar Signal Based on Multilevel Singular Value Decomposition and Savitzky-golay

雷达回波信号在接收过程中受到复杂噪声干扰,严重影响了目标检测的准确性,文中提出多级奇异值分解(multi-stage singular value decomposition,MS-SVD)和Savitzky-Golay(SG)的降噪方法,利用奇异谱分解的主成分分析特性,将代表噪声的奇异值逐级删减,对雷达信号的主分量进行重构,克服奇异值难以有效选取的难题,减少信号失真,再经SG平滑滤波处理,将局部强干扰剔除,实现雷达信号的提取.实验结果表明:该方法在各种信噪比条件下,能够有效地将雷达信号提取出来,在提高信号信噪比的同时,减少原始信号失真.

The radar echo signal is interfered by complex noise in the receiving process,which seriously affects the accuracy of target detection.In this paper,multilevel Singular Value Decomposition(SVD)and Savitzky-golay(SG)noise reduction methods were proposed to utilize the principal component analysis characteristics of singular spectrum decomposition.The singular value representing noise was cut down step by step,and the principal component of radar signal was reconstructed,which overcomes the difficulty of selecting singular value effectively and reduces signal distortion.After SG smoothing filtering,the local strong interference is eliminated,and the radar signal is extracted.Experimental results show that this method can effectively extract radar signals under various signal-to-noise ratio conditions,improve signal-to-noise ratio and reduce original signal distortion.