Skip to Main Content
A novel image fusion method based on discrete wavelet packet transform (DWPT) is proposed. First, after the registration of multi-source sensor images, the obtained images are decomposed into several sub-images of different frequency bands by using wavelet packet transform. Second, aiming at the characteristic of the sub-images after decompositions, fusion rules based on local region are chosen to fuse the coefficients of the sub-images, and a serial of fusion sub-images are obtained. Finally, the fusion image can be obtained by inverse discrete wavelet packet transform (IDWPT). By means of designing simulation experiments using visible and infrared image, the image fusion method based on wavelet packet transform is compared qualitatively and quantitatively with other five kinds of image fusion methods including weighted average, selecting maximum pixel gray value, selecting minimum pixel value, Laplacian pyramid transform and wavelet transform. Lots of experiments and comparisons with other fusion methods show that the effectiveness and superiority of image fusion method based on wavelet packet transform.