摘要
传统的小波变换阈值法去噪会使爆破振动信号的急剧变化部分产生人为的振荡现象,产生这种现象的原因是小波缺乏平移不变性。使用平移不变小波变换去噪方法,对爆破振动信号进行循环平移,利用软或硬阈值对该信号的小波系数进行压缩,重构信号,再进行相反的循环平移,通过多次的平移-消噪-平移,平均所获得的结果,从而消除小波基的平移依赖性。该方法不仅能有效地消除爆破振动信号的伪吉布斯(Psuedo-G ibbs)现象,而且能减小原始爆破振动信号和估计爆破振动信号之间的均方误差(MSE)。通过理论和实践证明了这种方法能够更有效地消除爆破振动信号中的噪声。
The traditional wavelet transform threshold de - nosing method may produce artifacts on the discontinuities of blasting vibration signals due to wavelet＇s lack of translation invariance. Translation invariant wavelet transform de - nosing method was used to perform the cycle translation for blasting vibration signals and the soft or hard threshold to compress and reconstruct of the wavelet coefficient of the signals. Then, the inverse cycle translation was conducted and the results of many times＇translation -de -nosing- translation were averaged so as to eliminate the translation dependence of wavelet base. This method can not only effectively suppress the Pseuedo - Gibbs phenomenon in blasting vibration signals, but also diminish mean square error （MSE） between the original blasting vibration signals and the estimated ones. It proves both theoretically and practically that this method can more effectively de - noise the blasting vibration signals.
出处
《金属矿山》
CAS
北大核心
2006年第2期12-14,21,共4页
Metal Mine
基金
国家"十五"科技攻关资助项目(2003BA612A-10-2)。
