基于NSCT和四邻域香农熵的多尺度图像融合算法

Multi-Scale Image Fusion Algorithm Based on NSCT and Four-Neighbor Shannon Entropy

针对多聚焦图像融合问题,提出一种非子采样Contourlet变换(non-sub sampled contourlet transform,NSCT)和四邻域香农熵的多尺度图像融合算法。算法特点是具有高边缘辨别特性的NSCT能够捕获来自不同父图像的所有显著目标边缘。对于配准的灰度图像,首先使用非线性各向异性扩散算法生成无噪声灰度图像,使用Contourlet变换获取所有无噪声图像的边缘,然后以四邻域熵算法生成熵图像,再采用winner-take-all方法生成最终的融合图像。实验结果表明:所提算法图像融合结果PSNR性能优于现有其他算法。

Aiming at the multi-focus image fusion problem,a multi-scale image fusion algorithm based on Non-Sub Sampled Contourlet Transform(NSCT)and four neighborhood Shannon entropy is proposed in this paper.The algorithm is characterized by an NSCT with high edge discrimination that captures all significant target edges from different parent images.For the registered grayscale image,the nonlinear anisotropic diffusion algorithm is first used to generate the noiseless grayscale image,the Contourlet transform is used to obtain the edges of all the noiseless images,then the 4 neighborhood entropy algorithm is used to generate the entropy image,and finally the winner-take-all method generates the final fused image.The experimental results show that the PSNR performance of the image fusion results of this algorithm is better than other existing algorithms.