联合小波和NLM滤波的侧扫声纳回波信号降噪  

Denoising method combining wavelet threshold transformation and NLM filtering for side-scan sonar signal

作　　者：王磊[1,2] 金绍华 崔杨 边刚 魏源 WANG Lei;JIN Shaohua;CUI Yang;BIAN Gang;WEI Yuan(Department of Military Oceanography and Hydrography,Dalian Naval Academy,Dalian 116018,China;92763 Troops,Dalian 116041,China;91937 Troops,Zhoushan 316002,China)

机构地区：[1]海军大连舰艇学院军事海洋与测绘系,辽宁大连116018 [2]92763部队,辽宁大连116041 [3]91937部队,浙江舟山316002

出　　处：《海洋测绘》2021年第3期69-73,共5页Hydrographic Surveying and Charting

基　　金：国家自然科学基金(41876103)。

摘　　要：为进一步降低侧扫声纳回波信号中非高斯分布的乘性噪声,获取更佳效果的侧扫声纳图像,提出了一种利用小波和NLM(nonlocal means)滤波的组合降噪方法。首先采用同态变换将侧扫声纳回波ping信号中的乘性噪声转换为加性噪声,然后利用小波阈值和NLM滤波对侧扫声纳每ping回波数据实施降噪处理,最后经过小波反变换和指数变换获取降噪后信号和图像。仿真实验和实测数据验证结果表明,该方法适用于侧扫声纳回波信号处理,可以获取较好的图像降噪效果。In order to reduce the non-gaussian multiplicative noise of side-scan sonar echo signal further and obtain better side-scan sonar images, a combined denoising method using wavelet threshold denoising and Nonlocal Means Filtering(NLM) is proposed in this paper. First, we convert the multiplicative noise in the side-scan sonar echo ping signal to additive noise with homomorphic filtering method. Second, we apply wavelet threshold transformation and NLM filtering method to process the single ping echo data of the side-scan sonar.The last, we obtain the denoised signal and image with the inverse wavelet transform and exponential transform.The simulation experiment and the measured data experiment show that the method of this paper is suitable for side-scan sonar echo signal processing, which can obtain better image denoising effect.

关 键 词：侧扫声纳 乘性噪声 同态变换 小波阈值降噪 NLM滤波 

分 类 号：P237.2[天文地球—摄影测量与遥感]