基于形态Haar小波变换的多聚焦图像融合

针对多聚焦图像融合问题，提出一种基于形态Haar小波分解和重构的新方法。通过形态Haar小波分解源图像，在低频分量中保留图像边缘和细节，并采用加权平均法进行融合。高频分量先经Gauss滤波去除噪声和边缘效应，再按取大值的原则进行融合。结合形态Haar小波重构融合后的高低频系数获得融合图像。实验结果表明，该方法能最大限度地保留图像边缘和细节信息，与总体平均法和小波变换法相比，融合图像的熵较大，总体交叉熵较小。

Multifocus Image Fusion Based on Morphological Haar Wavelet Transform

Aiming at multifocus image fusion, a new method is proposed based on morphological Haar wavelet decomposition and recomposition. By the new method, morphological Haar wavelet is used to decompose two original images, when image edges and details are hold in low frequency coefficients commendably. The mean of two low frequency coefficients is chosen as the low frequency coefficients of fused image. Gauss filtering is used to all high frequency sub-bands, wiping off noise and fringe effect. The bigger one of two corresponding high frequency coefficients is chosen as the same high frequency coefficients of fused image. Morphological Haar wavelet is used to recompose fused image from the fused coefficients. Experimental results show that this method performs better in preserving edge information and details for the test images than other image fusion methods.