摘要
为了避免常规正交小波变换所引起的图像边缘失真,特别采用了双正交小波变换进行图像的多分辨率分解。在融合处理时,为了获得更丰富的细节和更好的融合效果,对不同分解层、不同方向上的高频分量以及同一分解层图像中的不同区域均采用不同的融合算子。实验结果表明,该融合方法和融合规则是十分有效的,融合后图像的边缘和纹理细节都更加突出。
Multi-resolution decomposition of image is carried out with biorthogonal wavelet transform for avoiding image edge distortion caused by conventional orthogonal wavelet transform. In order to obtain richer detail and better fusion effect in image fusion, different fusion operators are used to different decomposition layers, high-frequency components in different directions and different regions of the same decomposition image layer. The experimental results show that the fusion method and fusion rule are very effective. The image edge and stripe detail can be more obvious after image fusion.
出处
《光电工程》
CAS
CSCD
北大核心
2004年第4期50-53,共4页
Opto-Electronic Engineering
基金
国防预研基金项目(51416050101DZ0138)
陕西省自然科学基金资助项目(J02141040)
关键词
图像融合
小波变换
图像分解
图像处理
Image fusion
Wavelet transform
Image decomposition
Image processing