活塞涡流信号的小波阈值降噪优化研究

Optimization of Piston Eddy Current Signal Based on Wavelet Threshold Denoising

涡流检测信号在产生、传输及接收过程中,会引入大量的随机噪声,为了后续信号特征提取识别处理,降噪在信号处理中是必不可少的一个环节。阐述了小波阈值降噪的基本原理,搭建了活塞喉口涡流检测实验平台。采集多组涡流检测信号,结合信噪比、均方根误差两项指标,评价降噪过程中阈值、阈值函数选取的不同对降噪效果的影响,对降噪方法进行比较优选。实验结果表明:当涡流检测信号中含有大量白噪声时,启发式阈值准则和改进指数型阈值函数能够很好的实现真实信号的提取,降噪效果最优。

In the process of generation, transmission and reception of eddy current detection signals, a large number of random noise will be introduced. Noise reduction is an essential part of signal processing for extracting and recognizing the characteristics of signals. The basic principle of wavelet threshold denoising was expounded, and the eddy current testing platform for piston throat was built. The effects of different thresholds and thresholding functions on signal de-noising are evaluated by collecting several sets of eddy current detection signals and combining signal-noise ratio (SNR) and root mean square error(RMSE). The experimental results show that when the eddy current detection signal contains large white noise, the ‘heursure’ criterion and the improved exponential threshold function can extract the real signal very well, and have better performance in noise reduction.