摘要
以描述小波系数分布规律的非高斯双参数模型为基础,介绍和分析了BiShrink滤波算法,指出了其存在的不足。提出的改进算法采用冗余小波变换替代正交小波变换,将子带内小波系数的局部相关性纳入滤波过程,给出了局部自适应的阈值估计策略,再通过双参数联合收缩函数达到系数收缩的目的。实验结果表明,改进算法同时兼顾了子带内小波系数之间的相关性和尺度间系数的传播特性,在有效滤除噪声的同时,较好地保持了图像的细节信息。
The paper presented an analysis of the BiShrink filtering algorithm based on the Non-Gaussian Bivariate M odels,and pointed out its disadvantages. It proposed an improved method,which adopted Redundant Wavelet transform instead of Orthogonal Wavelet transform,and brought the relativity among wavelet coefficients in the sub-band into the filtering procedure,then calculated the local self-adaption threshold based on a neighboring window,and finally acquired the estimated coefficient by the associated shrinkage function. The experimental results show ed that the new method gave consideration to the relativity among the wavelet coefficients and the transform characteristics of the inter-scale coefficients,which filtered the noise effectively and preserved the detail information of imagery.
出处
《海洋测绘》
CSCD
2015年第1期38-40,共3页
Hydrographic Surveying and Charting
基金
国家自然科学基金(60778051)