多小波阈值降噪法在钢丝绳缺陷检测中的应用

Application of Multiwavelet Threshold Denoising Method in Wire Rope Faults Detection

在介绍多小波理论和分析传统多小波阈值函数存在的问题的基础上,提出了一种新的多小波阈值函数.详细地讨论了该函数的数学性质,该函数对大于阈值的小波系数采取缓变的压缩策略,压缩量随着小波系数的增大而减小.将其应用于钢丝绳缺陷信号的降噪中,与传统多小波阈值函数的降噪进行了比较.实验结果表明,该阈值函数降噪后重建信号的信噪比(SNR)高,能检测到钢丝绳缺陷信号中的全部小奇异点,其降噪效果优于传统多小波阈值函数法.

On the basis of introduction of multiwavelet theory and analysis of existing problems of traditional multiwavelet threshold functions, a novel multiwavelet threshold function is proposed and its mathematical characteristics are discussed in detail. It adopts slow variation compression strategy for the wavelet coefficient which exceeds the threshold, the compression amount decreases as the wavelet coefficient increases. The improved function is applied to noise reduction of wire rope fault signals, and its effect is compared with that of the traditional multiwave threshold function. Experiment results indicate that, by using this threshold function, the reconstructed signal has a high Signal to Noise Ratio （SNR）, and that all the small singular points in wire rope fault signals can be detected. Thus the effectivity and superiority of the method are proved.