基于区域小波变换的序列显微图像融合  

Series microscopy images fusion based on area wavelet transform

作　　者：姜志国[1] 韩冬兵[1] 薛斌党[1] 周孝宽[1] 

机构地区：[1]北京航空航天大学宇航学院,北京100083

出　　处：《北京航空航天大学学报》2006年第4期399-403,共5页Journal of Beijing University of Aeronautics and Astronautics

基　　金：国家自然科学基金资助项目(60372017)

摘　　要：利用图像处理技术实现序列显微图像的融合显示是解决显微成像焦深有限的有效方法.针对空域融合算法在处理离焦步长较大和透明显微图像时存在的块效应和重影现象,基于图像像素间高度的相关性和小波变换多分辨率分析的特点,提出了一种基于区域小波变换的序列显微图像融合算法:在小波变换域内进行序列显微图像融合时,不同频段选择不同的融合算子;在确定融合的小波系数时,不仅考虑到相应位置的小波系数,还考虑了与它相邻的小波系数,使得融合后图像克服了空域算法存在的块效应和重影现象.同时,研究了不同的小波滤波器对图像融合质量的影响.实验证明,该方法具有很好的融合效果.Digital image fusion technique based on multi-focus series microscopy images is an effective method to extend the depth of focus. In order to avoid the blocking effect of fused image by spatial domain fusion algorithm when large focusing step is chosen or the image samples are transparent, a series microscopy images fusion algorithm based on area wavelet transform is proposed. Considering the fact that neighboring pixels are related and wavelet transform has the multi-resolution merit, fusion strategies are as follow： fusion operators are different corresponding to different frequency region, and fusion process is based on region coefficients of wavelet transform. Meanwhile, the influences of different wavelet bases on performance of fusion are investigated. Experimental results show that the blur is improved in obtained fused image and the blocking effect is eliminated. The fusion performance of this method is satisfactory.

关 键 词：图像融合 小波变换 显微图像 多层聚焦 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]