Various image fusion techniques have been conducted to merge various sensors data. However, when the existing image fusion processes are applied to the IKONOS images, the spectral distortion is appeared. To investigate this distortion problem, this work presents a study indicating that there are two reasons where the distortion of the spectrum is caused. To solve this color distortion problem, we proposed an image fusion process using Steepest Decent method which improves the correlation coefficients between the original 4m MS image and the fused 1m MS image. To evaluate our proposed image fusion process, the correlation coefficient between original MS (R, G and B) image and fused MS image and their average value have been compared for the correctness of spectral quality. Moreover the comparison among the proposed technique and existing processes (IHS image fusion, Brovey Transform, Principal Component Analysis) has been done. Our method has reduced the spectral distortion problem.
Here we proposed a process to minimize the spectral distortion in IKONOS 1m color image. In this method the 1m multispectral (MS) fused image is processed from 4m MS and 1m panchromatic (PAN) image to which the relations of spectral response between PAN and each band of MS image are obtained most. To maintain this relation, four parameters are added with the pixel value of each band of MS image and Steepest Decent method is used to update the pixel value of each band of fused MS image repeatedly until the color distortion becomes minimum with the maintenance of space information in PAN image as much as possible. Comparison among the proposed technique and existing four processes (IHS image fusion, Brovey Transform, Principal Component Analysis) has been done. Our purposed technique has succeeded to generate the 1m fused image where space information on PAN and MS spectrum information were maintained highly and spectral distortion has been reduced significantly compared to these existing methods.
For image fusion applications, the critical issue is how to produce a fused imagery with minimizing color distortion and maximizing the crease. However, due to nonsymmetrical spectral responses of IKONOS imagery, most existing techniques produce color distortion in 1m fused images. In our another report we have succeeded to reduce this color distortion significantly by using four spectral weighting parameters in our proposed fusion technique [1]. However, block distortion appeared in part of the 1m fused image. In this work we investigated the causes and proposed a sharpening process which uses Wavelet Transform to minimize this block distortion in 1m fused MS image. We have compared the 1m fused image before and after this sharpening process and found that the block distortion has almost been disappeared. Moreover, distortion of color of the entire image did not changed after this process.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.