机构地区: 华中科技大学机械科学与工程学院
出 处: 《振动工程学报》 2008年第5期521-529,共9页
摘 要: 利用改进的Coiflet设计算法,得到了阶为12,14,16,18,20的高阶Coiflet滤波器系数。通过与Daubechies构造的低阶Coiflet滤波器系数的比较,改进算法的计算结果更精确。对高阶Coiflet滤波器的相频特性进行了计算和验证,结果表明高阶Coiflet的尺度和小波函数图形与低阶基本相似。小波滤波器线性相位特性比尺度滤波器好。尺度滤波器相位线性差,而小波滤波器相位近似线性。随着Coiflet阶数的提高,小波滤波器线性相位变化更快。尺度滤波器相位在低频段近似为0。给出了5个新的高阶和具有更高精度的5个低阶Coiflet滤波器系数表,同时还给出了构造高阶区间Coiflet尺度函数的应用。 Higher order Coiflets for which both the scaling and wavelet functions have a high number of vanishing moments are constructed by the modified Daubeehies method. In contrast with Daubechies' filters of lower order Coiflets, our results are more accurate. The properties of phase and vanishing moment of the scaling and wavelet functions and the corresponding filter coefficients are calculated and verified. The results show that the figures of higher order Coiflets are similar to that of lower order ones. The wavelet filter is superior to the scaling filter in linear phase. The phase of the scaling filter is non-linear, while the phase of the wavelet filter is almost linear. With the increase of the order of Coiflet, the linear phase of the wavelet filter changes sharply. The phase of the scaling function is almost zero in the low frequency band. A filter coefficient table of 5 higher order Coiflets and 5 original lower order coiflets with high precision is listed. An application of higher order Coiflet to intervallic one is presented.