应用小波多分辨分解及重构技术和小波包分析技术进行检测信号预处理(噪声和干扰信号分离与去除)的基本原理,结合电涡流扫描检测实验,采用离散小波变换强制消噪法、软阈值消噪法和不同熵准则的小波包分析消噪法对检测信号进行预处理,并以SNR和RMSE为判断消噪效果好坏的标准,进行了效果的比较和优选。从理论分析和实验研究结果可知,分离提离等干扰信号,可采用强制法;消除高频噪声,采用基于Shannon熵准则的WPA法,效果最好。
Eddy current signal pre-processing techniques for noise elimination using the discrete wavelet transform and wavelet packet analysis methods are studied. Combining the multi-layered conductive structures defect inspection experiments, the signal processing methods presented are applied and results are compared with the SNR and RMSE as criteria and the best de-noising method is recommended.