ITD改进信号子空间超声检测信号去噪  

Improved signal subspace method of ultrasonic test signal denoising based on ITD

作　　者：李大中[1] 赵杰[1] 刘建屏[2] 蔡文河[2] 马延会[2] 

机构地区：[1]华北电力大学自动化系,河北保定071003 [2]华北电力科学研究院有限责任公司,北京100045

出　　处：《中国测试》2016年第4期102-106,共5页China Measurement & Test

摘　　要：金属材料超声检测信号中存在的干扰噪声严重影响实际缺陷的检测精度,因此必须对采集的检测信号进行有效去噪处理。针对传统信号子空间去噪方法的不足,提出固有时间尺度分解(ITD)改进信号子空间信号增强算法的超声检测信号去噪方法,融合ITD方法和信号子空间的优势。通过现场超声检测缺陷信号数据的对比验证表明,信噪比、均方根误差和相关性等参数都比较满意,对含有有色噪声的超声检测信号去噪效果明显。Interfering noise in ultrasonic testing signals of steel materials has seriously affected the testing precision of actual defects. Therefore, the noise in collected testing signals must be eliminated. A noise-removing method is proposed in this paper to solve the disadvantages of traditional signal subspace based on intrinsic time-scale decomposition（ITD） improved signal subspace. It combines the strengths of the ITD method and the signal subspace. SNR,RMSE and correlation as well as other parameters are satisfactory and the colored noises of ultrasonic defect signals are largely reduced,according to the on-site contrast verification of ultrasonic testing signal data.

关 键 词：超声检测 去噪 固有时间尺度分解 信号子空间 

分 类 号：TN911.7[电子电信—通信与信息系统]