基于高频方差对比度小波变换的图像融合方法

An Image Fusion Method Based on High Frequency Variance Contrast Processing of Wavelet Coefficient

针对现有的区域图像融合方法没有充分利用代表图像细节信息的高频分量的情况,给出一种以高频方差对比度为判断依据的小波变换的图像融合算法。算法以高频域为研究对象,给出了高频方差对比度的概念,其步骤为:首先利用小波变换得到待融合图像的多分辨分析,再以相应各级上的对应区域的高频方差对比度为判据,选择图像的多分辨分析的相应各级上的小波系数,最后通过小波重构得到融合图像。实验结果性能分析表明,充分发挥了高频分量的作用,能使融合图像综合更多的源图像信息,融合效果良好。

The high frequency components representing image details have not been fully utilized in many current regional image fusion methods. In view of this fact, a variance contrast - based image fusion algorithm with wavelet transform is presented for fusing registered images. This wavelet transform image fusion algorithm uses high - frequency variance contrast as the basis for judgement. The algorithm is for high frequency in this paper and the defini- tion of the high frequency variance - contrast is presented. Its steps are ： firstly, the input images are decomposed by wavelet transform ; secondly, variance - contrast is adopted for fusing high frequency coefficient ; finally, the fusion image is recontrueted by wavelet inverse transform. The result shows that this algorithm can fully utilize the specialty of high frequency components and preserve more imformation from premitive images.