基于Contourlet变换的多聚焦图像融合算法研究

针对传统图像融合算法某些方面的不足,提出了一种基于Conlourlet变换的多聚焦图像融合算法。首先对两幅多聚焦图像进行Contourlet变换,得到不同分辨率层次下的Contourlet系数,包括低频分量和高频分量,利用Contourlet变换将图像的低频子带与高频子带进行分离,低频部分采用平均方法求出相关系数,高频部分采用绝对值取大的方法求出相关系数,对两个相关系数进行融合处理。实验结果表明,与传统的多聚焦图像融合算法相比,该算法增强了图像融合在细节方面的表现能力。

Study on multi-focus image fusion algorithm based on Contourlet transform

In light of some shortcomings of the traditional image fusion algorithm,the multi-focus image fusion algorithm based on contourlet transform is put forword. Firstly,contourlet transform is applied to the two multi-focus images to get the contourlet coefficient under different resolution levels,including low frequency and high frequency components. Using Contourlet transform,image,low frequency band and high frequency band were separated. For low frequency component,average fusion algorithm is used to obtain the correlation coefficient and for high frequency components,maximum absolute value selection is used to get the correlation coefficient. Finally,The two correlational coefficients are handled with fusion. The experimental results show that compared with the traditional multi-focus image fusion algorithms,this algorithm can enhance better fusion results at micro levels.