结合NSCT和区域特征的图像融合新算法

提出一种把非下采样Contourlet变换(NSCT)和区域特征相结合的图像融合新方法.该方法能够获取更好的空域和频域中的局部特征,同时提高融合图像的质量.用NSCT对已经配准的源图像在不同尺度和方向进行分解,低频子带分量采用区域平均能量和匹配度相结合的融合规则,高频子带分量使用改进的拉普拉斯能量和取大的融合规则.然后,利用逆NSCT变换对图像重构得到融合结果.实验结果表明,新方法优于其他三个常用的方法,且较好地保留图像的边缘和细节信息.

A novel image fusion algorithm based on NSCT and regional features

The wavelet transform fusion method has its limitations.To solve this problem,a novel image fusion algorithm is proposed based on Nonsubsampled Contourlet transform (NSCT) and regional features.This method can obtain better the local features of the spatial and frequency domain,while improving the quality of the fusion image. Firstly, two primitive matched images are decomposed at multi-scale and multi-direction by NSCT. Then the fusion rule based on regional average energy and the average energy matching degree is used to fuse the low frequency subband components.The high frequency subband components by using improved sum of Laplace energy. Finally, the fusion result of image reconstruction is obtained through inverse NSCT.The results of simulation show that this novel method are better than the other three common image fusion methods.The test results can preserve a good edge and many detail information.