基于非下采样剪切波变换与显著信息加权的图像融合算法

为了克服当前较多图像融合方法主要依靠测量图像能量信息来完成不同系数的融合,忽略了图像的显著内容,导致融合图像含有吉布斯效应及间断效应等弊端,设计了非下采样剪切波变换(nonsubsampled Shearlet transform,NSST)耦合显著信息加权的图像融合算法.引入NSST机制,对源图像进行系数分解,获取高、低频系数.借助高斯滤波器来构造出显著度量模型,以计算图像拥有的显著信息.随后,利用信息熵函数来计算出图像拥有的细节丰富度.并以图像拥有的细节丰富度和显著信息为依据,设计低频系数融合的加权因子,以此完成低频像素的融合.最后,利用图像中像素点的三邻点像素值,融合高频系数,获取融合图像.实验结果显示,与当前图像融合技术相比,所提算法融合质量更好,融合结果连续性较强,所对应的平均梯度值较大.

A Image Fusion Algorithm Based on Nonsubsampled Shearlet Transform Coupled with Significant Information Weighting

In order to overcome the shortcomings of the current image fusion methods, which mainly rely on the measurement of image energy information to complete the fusion of different image coefficients, ignoring the significant information of the image, resulting in the Gibbs effect and discontinuity effect of the fused image, this paper designs an image fusion algorithm using the significant information weighting based on the non lower sampling shear wave transformation. In order to obtain high and low image coefficients, the better translation characteristic is introduced to decompose the source image coefficients. The saliency measurement model is constructed by Gaussian filter, and the saliency information of image is calculated. Using the information entropy model to calculate the information richness of the image. Based on the information richness and significant information of the image, the weighted factor of low-frequency coefficient fusion is formed and low-frequency coefficient is fused. The fusion function of high-frequency coefficients is constructed by using the three adjacent pixel values of pixels in the image. The experimental results show that compared with the current image fusion algorithm, the fusion performance of this algorithm is better, and the image fusion has fewer disadvantages and higher quality.