一种基于多尺度最大信息熵和梯度的图像融合算法

An ImageFusion Algorithm Based on Max Comentropy and Gradient of Multiple Scales

提出了一种多尺度最大信息熵(Max Information Entropy,MIE)及梯度的图像融合算法,该算法在对源图像多尺度分解的基础上,根据低频小波系数及高频小波系数的特点,把信息熵引用到小波低频系数的选择中,根据局部信息熵的大小确定小波系数的选择;而高频采用基于最大梯度值的方法,最后对所选小波系数进行重构,即可得到融合图像。二者的结合,对图像的细节处理更加细致,又有效地消除冗余信息。通过实验分析,结果表明该算法与其他基于区域的方法相比,提高了融合效果。

In this paper, a new image fusion algorithm based on multiple scales and MIE and gradient is presented. The information entropy is used in the algorithm, according to the size of the local information entropy to determine the choice of wavelet coefficients; and high - frequency maximum gradient value based method, and then reconstruction of the selected wavelet coefficients. Combination of the two pairs of the details of the image processing is more detailed, but also effectively eliminate the redundant information. Through experimental analysis, the results show that the algorithm improve the fusion effect compared with other fusion algorithm based on region.