基于盲源分离的超声信号去噪的仿真研究  

Computer simulation study on noise cancellation for ultrasonic signals based on blind source separation

作　　者：冯宪高[1] 万江飞[2] 郭晓疆[3] 刘清坤[4] 李维平[1] 

机构地区：[1]中国石油天然气管道第四工程公司,河北廊坊065000 [2]中国石油天然气集团公司通信公司,河北廊坊065000 [3]中国石油天然气管道科学研究院,河北廊坊065000 [4]上海交通大学自动检测研究所,上海200240

出　　处：《传感器与微系统》2006年第7期36-39,共4页Transducer and Microsystem Technologies

摘　　要：提出一种新的基于盲源分离的超声信号去噪方法。为了验证去噪方法的有效性,应用此方法处理了仿真的超声信号,并与小波去噪的效果进行了比较。实验结果表明:该去噪方法能极大提高超声信号的信噪比,且其效果能与小波去噪方法相媲美,其特点是通过超声信号和噪声信号的盲源分离实现噪声消除。A new denoising method for ultrasonic signal based on blind source separation（BSS） is proposed. It is applied to denoising the simulated ultrasonic signals processing and compared with wavelet denoising method. Experimental results show that signal noise ratio of ultrasonic signal can be improved greatly and its denoising performance is very near to wavelet denoising method. The denoising method is realized by blind source separation of the ultrasonic signal and noise.

关 键 词：超声检测 无损评估 盲源分离 独立分量分析 

分 类 号：TG115.28[金属学及工艺—物理冶金]