基于DT-CWT统计模型的舰船噪声信号中线谱信号检测研究

双树复解析小波变换(DT-CWT)在信号去噪方面的性能优于实小波变换,且计算量远小于后者.文章基于DT-CWT小波理论,通过对海洋环境噪声中舰船噪声中低频线谱信号小波系数的层间联合分布的分析,提出一种DT-CWT统计模型并推导出最大后验概率估计子(MAP),用于检测海洋噪声背景中的舰船噪声中的低频线谱信号.对实测舰船噪声信号和海洋环境噪声的分析表明,所提出的DT-CWT统计模型算法明显优于VisuShrink、SureShrink和BayesShrink算法对舰船噪声中线谱信号的检测效果.

Effectively Detecting Line Spectrum of Ship-Radiated Noise Using DT-CWT Statistical Model

In the introduction of the full paper,we point out that the low-frequency components of the line spectrum of ship-radiated noise are strongly non-Gaussian.This,we believe,is very important;around it,we propose our effective method of detection.Section 1 of the full paper,in addition to briefing the DT-CWT(Dual-Tree Complex Wavelet Transform)statistical model,derives eq.(9)for computing the MAP(maximum a posteriori)estimator.Section 2 gives a four-step procedure of our signal detection algorithm;in step 3,eq.(9)is used.Subsection 3.2 presents in Figs.8 and 9 the detection results of ship-radiated noise obtained by four different algorithms:VisuShrink,SureShrink,BayesShrink,and the DT-CWT statistical model.These results show preliminarily that the DT-CWT statistical model algorithm proposed by us is more effective than the other three algorithms.