基于有限脊波变换的图像融合方法

Novel algorithm of multi-sensor image fusion using FRIT

针对小波变换在进行图像处理时不能有效表示直线/曲线的奇异性问题,研究了能够很好表示二维或更高维奇异性的有限脊波变换的基本原理,讨论了将有限脊波变换应用于图像融合的可行性,提出了基于有限脊波变换方法进行图像融合的基本架构,并对变换系数的设置与融合过程的处理进行了详细说明。仿真实验的主客观性能比较表明,基于有限脊波变换的图像融合方法比基于金字塔变换和基于小波变换方法的融合效果要好得多。

The problem that the contour segment of 2-dimensional image could not be gotten perfectly with wavelet transform in image processing is analyzed. The linear and curve singularity could not be expressed by wavelet transform. Then, the principal of finite ridgelet transform and its good performance in expressing the singularity of 2-dimensional or high dimensional are studied. The feasibility of image fusion using finite ridgelet transform is discussed in detail. A new fusion method based on finite ridgelet transform and its fusion frame- work are proposed. The coefficients structure and the fusion procedure are given in detail. Experiments show that the proposed algorithm works better in preserving the edge and texture information than other methods do in image fusion.