摘要
在基于小波分解的图像融合中,低频系数的选取决定融合图像的轮廓,同时影响图像的边缘细节。双匹配度图像融合算法首先在区域能量的基础上合理引入区域梯度,然后根据匹配度的类型和测量值进行小波低频系数的选取。仿真实验从两种无需基准融合图像的客观评价方法和视觉效果两个方面表明双匹配度融合算法是一种有效的图像融合算法。该算法能够充分地利用区域特征信息,有效地保留图像的轮廓和边缘信息,同时能够有效地避免块效应的产生,产生良好的视觉效果,且优于一些经典的图像融合算法。
For image fusion using 2D discrete wavelet transform （DWT）, the low frequency coefficients decide the contour of the fused image and affect on the edge details of the fused image. Image fusion algorithm based on double match measure is performed by introducing the local gradient into the local energy and then selecting of the low frequency coefficients according to the type and value of match measures. The results from two objective evaluation methods without reference fusion image and the visual show that the image fusion algorithm based on double match measure is effective, in which the local features are fully applied, and the contour and edge details of the image fusion are efficiently retained, so the block effect can be effectively avoided, and a good visual effect is produced, it is better than some traditional image fusion algorithms.
出处
《红外技术》
CSCD
北大核心
2006年第12期686-690,共5页
Infrared Technology
基金
国家自然科学基金(60402038)
国家重点基础研究发展计划973(2006CB705700)
关键词
双匹配度
多光谱
图像融合
小波变换
double match measure
multi-spectral
image fusion
wavelet transform
