基于非因果信号处理的分数阶次低通滤波器

摘要

0就目前而言,对于低通滤波器都不能在去除噪声与保留图像的边缘和纹理细节之间达到权衡.针对上述情况,进一步分析并研究分数阶次积分的幅频与相频的特性,再结合非因果信号处理思想,从而提出一种基于非因果信号处理的非因果分数阶次低通滤波器NCFLPF(Non-causal Fractional Low-pass Filter).通过在合成和自然噪声图像上的定量和定性对比实验表明,所提新方法不仅能抑制滤波过程中的相位失真,而且还能在有效平滑噪声的同时,尽量保留着图像的细节.%Since the existing low-pass filters failed to reserve the edge and texture detail while eliminating the image noise,we analysed and studied further about the characteristics of the amplitude-frequency and the phase-frequency of forward and backward fractional integration.Then,combined with the non-causal signal processing,a novel non-causal fractional low-pass filter (NCFLPF) based on non-causal signal processing is proposed.The qualitative and quantitative experiment on synthetic and natural noise image indicated that the proposed method not only suppress the phase distortion,but also reserve the image detail while effectively eliminate the noise.