Abstract: The higher density discrete wavelet transform with two generators can achieve intermediate scales analysis and is ap-proximately shift invariant, so that it is very suitable for signal denoising. Aiming at the bad symmetry of the current higher density wavelet bases, an approach for the construction of the least asymmetric higher density wavelet bases is proposed. The proposed method optimiza-tionaly selects the nearest linear-phase low-pass filter from all the possible factorization results of the maximally flat linear-phase FIR filter, subsequently computes the two wavelet filters, and then the least asymmetric higher density wavelets can be constructed. Simulation ex-periments show that using the new wavelets can improve the denoising performance of the higher density discrete wavelet transform com-pared with the current wavelets. The results of the application to denoising a roller bearing fault vibration signal show that the constructed wavelet can effectively extract the bearing fault characteristic from that signal.
秦 毅;汤宝平;王家序. 最小不对称高密度小波基的构造及应用[J]. , 2010, 29(10): 195-200.
QIN Yi;TANG Bao-ping;WANG Jia-xu. The construction of the least asymmetric higher density wavelet bases and its application6885. , 2010, 29(10): 195-200.