基于纹理和梯度特征的多尺度图像融合方法

Multi-scale image fusion method based on the characteristics of texture and gradient

该文提出一种基于纹理和梯度特征的多尺度图像融合方法。采用纹理提取滤波器和边缘梯度滤波器模板分别对每层Gauss金字塔进行滤波,生成多尺度纹理和边缘图像,使得多分辨率变换域中能够将原图像中的纹理信息包含进来,从而为进一步融合提供更全面的信息量度。采用该文所提出的方法对红外与可见光图像进行融合仿真,并且与具有代表性的融合方法进行了比较。实验结果与图像融合质量评价显示了该文所提出方法的优越性。

This paper presents a multi-scale image fusion method based on texture and edge.After each level of Gaussian Pyramid was separately filtered by texture-extracting filters and edge gradient filters,a series of texture and edge images were generated including the texture information of the original image,which provides more comprehensive information measurement for further fusion.Infrared and visible images were fused utilizing the proposed image fusion method,with the fusion results compared with those of several traditional fusion methods.Experimental results and image fusion quality assessment show that the proposed fusion method demonstrates more superiority.