具有边缘保持特性的零树小波图像压缩算法  

ZERO-TREE WAVELET IMAGE COMPRESSION ALGORITHM WITH EDGE PRESERVING CHARACTER

作　　者：田迎华[1] 崔广才[1] 

机构地区：[1]长春理工大学计算机科学技术学院,吉林长春130022

出　　处：《计算机应用与软件》2009年第1期80-81,95,共3页Computer Applications and Software

基　　金：吉林省科技发展计划基金项目(20060529)

摘　　要：基于小波变换的图像压缩算法,在含噪图像和较低码率时出现的边缘模糊现象多年来一直未能得到很好的解决。为了解决这一问题,提出一种具有边缘保持特性的零树小波图像压缩方法。首先,对图像进行小波边缘检测,确定哪些小波系数是图像的边缘特征,将其保护起来;然后,对小波变换域系数采用改进软阈值收缩方法实现去噪;最后,利用SPIHT(Set Partitioning in Hier-archical Trees)算法对图像进行压缩编码。实验结果表明,该方法不仅能获得较高的图像压缩率、较好地去除噪声,而且能在一定程度上解决边缘模糊问题,能较高地恢复图像质量。Edge blurring phenomenon, which occurs in the wavelet-based image compression algorithms under noised image or low bit rates,remains an open question for many years. In this paper it presents a zero-tree wavelet image compression algorithm with edge preserving character to solve this problem. Firstly, the image is wavelet edge detected, the determined wavelet coefficients which are of the edge character of image will be protected,then denoise the wavelet transform domain coefficients by improved soft -threshold contraction method ; and at last the image is compressed and coded by Set Partitioning in Hierarchical Trees （SPIHT） method. The experimental result shows that this method achieves high compressing ratio, removes noise well,and also solves edge blurring problem to an extent with higher quality for recovered image.

关 键 词：图像处理 压缩编码 边缘检测 小波阈值去噪 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] TN919.81[自动化与计算机技术—控制科学与工程]