基于拼接思想的小波域多聚焦图像融合

A wavelet domain multi-focus image fusion based on mosaic

分析多聚焦图像融合的特点,给出一种基于拼接思想的小波域图像融合方法。利用小波分解系数中区域方差和较大的系数对应着较好的图像块原则,结合区域一致性检验修正,选择代表优质图像质量的小波系数,经小波反变换重构聚焦图像。实验结果表明:采用此方法得到的融合图像清晰度高,图像特性保留良好,客观评价指标有明显提升。

Through analyzing the features of multi-focus image fusion,a novel mosaic-based image fusion algorithm was proposed in wavelet domain.The fused image is rebuilt from the wavelet coefficients corresponding to better quality blocks in the original images,which are selected according to the criterion of maximum sum of coefficients variance in each block and corrected with consistency.The experimental results show that the quality of fused image has a significant improvement in subjectivity as well as the objective measurements under a variety of different principals.