基于快速有限剪切波变换的图像去噪  

Image Denoising based on Fast Finite Shearlet Transform

作　　者：荆方 刘增力[1] JING Fang;LIU Zeng-li(Information Engineering and Automation College,Kunming University of Science and Technology,Kunming Yunnan 650500,China)

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《通信技术》2019年第2期323-329,共7页Communications Technology

基　　金：基于现代信号处理的高速列车辐射噪声分析(No.61271007)~~

摘　　要：图像去噪一直是数字图像处理领域研究的热点之一。近年来,针对小波变换的不足,提出了一种基于变换域的多尺度几何理论分析——剪切波变换(Shearlet Transform)。结合剪切波变换的特点,提出了一种基于快速有限剪切波变换(Fast Finite Shearlet Transform)的图像阈值去噪方法。在噪声强度相同的情况下采用同样的阈值去噪算法,将快速有限剪切波变换与传统的小波变换进行对比。仿真结果表明,有限离散剪切波变换去噪算法相比于小波变换来说,信息冗余量低,去噪后图像的失真较少,且去噪耗费的时间较短,在峰值信噪比、冗余及去噪所耗费的时间等方面都优于小波变换。Image denoising is always one of the hot topics in the field of digital image processing.In recent years,in view of the shortcomings of wavelet transform,a multi-scale geometric theory analysis based on transform domain-shearlet transform is proposed.Combined with the characteristics of shearlet transform,an image threshold denoising method based on fast finite shearlet transform is proposed.Under the same condition of noise intensity,the same threshold denoising algorithm is used to compare the fast finite shearlet transform with the traditional wavelet transform.The simulation results show that the finite discrete shear wave transform denoising algorithm has lower information redundancy than wavelet transform,and the distortion of the image after denoising is less,and the time for denoising is shorter,and thus is obviously superior to wavelet transform in terms of peak SNR,redundancy,and time spent on denoising.

关 键 词：图像处理 多尺度几何分析 剪切波变换 小波变换 阈值去噪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]