一种改进的基于全变差范数图像融合方法

针对基于全变差范数图像融合方法在多聚焦图像融合中产生分块效应的缺陷，提出将区域的特征表示与主成分分析思想相结合的改进算法，解决了单独块变换造成的图像分块效应问题.实验结果表明，改进算法能有效地消除分块效应，且针对含噪图像，其融合效果明显优于离散小波方法.

An Improved Total Variation-Based Image Fusion Algorithm

Considering the total variation-based algorithm for image fusion has block effects when fusing multi-focus images, the notion of feature representation in conjunction with principal component analysis was used for improvement. The improved approach prevents independent block transform from happening. The experimental results show that the proposed algorithm can resolve the problem of block effects efficiently, and it also outperforms the discrete wavelet transform based algorithm when focusing on noisy image. Key words：