摘要
针对传统的硬软阈值函数和一些改进的典型小波阈值函数,没有同时解决函数不连续、与原始系数存在恒定偏差和在阈值处不可导,导致去噪后图像出现振铃、伪吉布斯现象和失真等问题,提出一种改进的阈值函数。使用Matlab进行仿真,对高频细节图像和低频细节图像进行去噪实验,实验结果说明,改进的阈值函数较硬软阈值函数和其余几种阈值函数去噪效果更好。提出的改进阈值函数更加光滑,在图像小波去噪中有着更好的应用前景。
The traditional hard and soft threshold function and some improved common wavelet threshold functions are discontinuous,have constant deviation between the estimated coefficients and the original coeffients and have no derivation at the threshold,which leads to ringing and pseudo-Gibbs phenomenon and distortion in the denoising image.And an improved threshold function is proposed.Matlab software is used to complete the denoising simulation experiments for images having high-frequency and low-frequency details.The experimental results show that the proposed threshold function is better than the hard and soft threshold function and the other common threshold function in image denoising.The proposed improved threshold function is smoother and has a better application prospect in image wavelet denoising.
作者
黎彪
刁燕
罗华
徐明
LI Biao;DIAO Yan;LUO Hua;XU Ming(School of Mechanical Engineering,Sichuan University,Chengdu 610065)
出处
《计算机与数字工程》
2020年第6期1460-1463,1506,共5页
Computer & Digital Engineering
基金
四川大学泸州市人民政府战略合作项目(编号:2018CDLZ-22)资助。
关键词
图像去噪
小波变换
阈值函数
image denoising
wavelet transform
threshold function
