Image Fusion Using Balanced Multiwavelet Transform

Mg Xia,Y He,F Su,Rj Yang,Y Chai
2004-01-01
Abstract:A novel fusion method for optical image and Synthetic Aperture Radar (SAR) image using balanced multiwavelet transform is proposed to fuse optical image and Synthetic Aperture Radar (SAR) image, and the multiwavelet function of the orthogonal balanced multiwavelet constructed are symmetric and asymmetric, respectively. Support set of the constructed balanced multiwavelet is 3. Its equivalence scalar low-pass filter banks have full. low-pass property. In this image fusion scheme,decomposition depth of source images is 3 levels using balanced multiwavelet transform. The composite coefficients are obtained through a sample-based maximum selection rule. In order to get better description of the scene, fused multiscale image is enhanced in wavelet transform domain. The larger entropy of a image is, the more information it contains. This image fusion algorithm is compared quantitatively with Db4 scalar wavelet's, GHM multiwavelet's, and CL4 multiwavelet's using entropy as a performance evaluation criteria of image fusion. Experimental results demonstrate that the balanced multiwavelet image fusion algorithm which increments entropy max outperforms other multiwavelets-based and scalar wavelet-based fusion algorithms. Fused image using. balanced multiwavelet is suitable for human visual system and image edges can be preserved well using. this image fusion algorithm.
What problem does this paper attempt to address?