基于小波系数模糊积分的图像融合算法研究

Image fusion algorithm based on fuzzy integral

提出一种新的基于Choquet模糊积分的图像融合方法。首先利用小波系数的区域特征获取模糊积分的信度函数,其次根据局部窗口内的模糊边缘评价函数自适应地构造模糊密度,最后由Choquet模糊积分确定融合后的小波系数,从而得到融合图像。将此方法应用于微光与红外图像融合处理中,实验结果表明,Choquet模糊积分法可以使得最终的融合图像在空间细节信息的增强和光谱信息的保持两方面的综合性能得到提高。

A novel image fusion algorithm was proposed, which was based on choquet fuzzy integral. The main idea was to obtain belief function using the wavelet coefficient of infrared images and low light level images, then construct fuzzy density using the fuzzy edge evaluation function of local area, finally get fused wavelet coefficient by fuzzy integral and reconstruct fused image by inverse wavelet transform. The experimental results show that the method not only keeps spectral quality of fused image, but also enhances the spatial detail quality of fused image.