基于多形态学成分分析的图像融合

Image Fusion Based on Multi-morphological Component Analysis

摘要:

将多尺度分解与稀疏表示相结合，提出了一种基于多形态学成分分析（MCA）的图像融合算法。采用基于联合稀疏表示（JSR）的方法融合卡通子图像中的冗余和互补信息，并利用基于方向特征的方法融合具有更多细节信息和噪声的纹理子图像。结果表明，提出的图像融合算法在主观视觉效果和客观评价指标上均优于先进的图像融合算法。

Abstract:

By combining the multi-scale decomposition and sparse representation， an image fusion algorithm based on multi-morphological component analysis （MCA） is proposed in this paper. The fusion method based on joint sparse representation （JSR） is employed to fuse the redundant and complementary information in the cartoon sub-images， and the fusion method based on directional feature is used to fuse the texture sub-images with more detailed information and noise. The results show that the proposed algorithm is superior to the state-of-the-art image fusion methods in subjective visual effects and objective evaluation metrics.