Pansharpening optimisation using multiresolution analysis and sparse representation. (3rd July 2017)

• Record Type: Journal Article
• Authors: Imani, Maryam; Ghassemian, Hassan
• Is Part Of: International journal of image and data fusion. Volume 8:Number 3(2017)
• Page Start: 270
• Page End: 292
• Publication Date: 2017-07-03
• Subjects: Convex objective function -- multiresolution -- sparse representation -- image fusion; Image processing -- Periodicals; Multisensor data fusion -- Periodicals; Multisensor data fusion; Periodicals; 621.36705
• DOI: 10.1080/19479832.2017.1334713
• Languages: English
• ISSNs: 1947-9832

Abstract

ABSTRACT: The fusion of multispectral (MS) image containing high spectral resolution with panchromatic (PAN) image containing high spatial resolution is called pansharpening. It is a major issue of research in remote sensing image processing. In the proposed fusion method, a convex objective function is defined that involves the desirable characteristics such as preserving the spatial structure of PAN and the spectral information of MS in the fused product. The similarity coefficients related to a linear relationship between MS, PAN and the fused image are utilised in the defined objective function. These similarity coefficients are used to transfer both the spatial and the spectral information with high resolution to the fused image. The coefficients are learnt by the proposed fusion approach based on multiresolution analysis and sparse representation. In addition to fuse MS and PAN images, the proposed method is applicable for fusing other source images such as multi-focus, visible-infrared and medical images. The experimental results show the advantages of the proposed fusion method compared to some popular and state-of-the-art fusion methods using quantitative and qualitative assessments.