基于分块信息熵的彩色图像融合算法

针对常用的彩色图像融合算法在融合过程中降低原图像空间分辨率的问题，提出了一种基于分块信息熵的彩色图像融合算法。算法通过对图像进行块划分，计算分块图像在 IHS 色彩空间中亮度分量 I 上的信息熵，以对应块中信息熵大的为主元进行融合，得到新的融合图像。实验表明，与常用的IHS变换及PCA图像融合算法相比本文提出的方法可以更好的统筹源图像的场景细节信息，且在信息熵与平均梯度等方面都有很好的表现，具备一定的实用价值。

Color Image Fusion Algorithm Based on Block Information Entropy

A method of color image fusion is proposed based on block information entropy, concerning the spatial resolution reduction of the original artwork during the fusion process with the traditional image fusion method. The developed method generates new fusion image by blocking images, calculating the information entropy of the blocked image on the luminance component I in color space IHS, and fusing the pivot element of the corresponding block whose information entropy is more. The experiment suggests that the method proposed in this thesis is better capable of planning scenic detailed information as a whole, compared to the traditional color image fusion method. In addition, it performs exceptionally in many aspects including information entropy and average gradient. Therefore, it is indeed of some practical value.