Variational Bayesian Pansharpening with Super-Gaussian Sparse Image Priors.

Fernando Pérez-Bueno,Javier Mateos,Miguel Vega,Rafael Molina,Aggelos K Katsaggelos

doi:10.3390/s20185308

Fernando Pérez-Bueno, Javier Mateos + Show 3 more

Open Access

https://doi.org/10.3390/s20185308

Copy DOI

Abstract

Pansharpening is a technique that fuses a low spatial resolution multispectral image and a high spatial resolution panchromatic one to obtain a multispectral image with the spatial resolution of the latter while preserving the spectral information of the multispectral image. In this paper we propose a variational Bayesian methodology for pansharpening. The proposed methodology uses the sensor characteristics to model the observation process and Super-Gaussian sparse image priors on the expected characteristics of the pansharpened image. The pansharpened image, as well as all model and variational parameters, are estimated within the proposed methodology. Using real and synthetic data, the quality of the pansharpened images is assessed both visually and quantitatively and compared with other pansharpening methods. Theoretical and experimental results demonstrate the effectiveness, efficiency, and flexibility of the proposed formulation.

Highlights

Remote sensing sensors simultaneously capture a Multispectral (MS) low resolution image along with a single-band high resolution image of the same area, referred to as Panchromatic (PAN) image.MS high-resolution images are needed by many applications, such as land use and land cover analyses or change detection
Pansharpening is a technique that fuses the MS and PAN images into an MS high resolution image that has the spatial resolution of the PAN image and the spectral resolution of the MS one
We model the relation between the MS high resolution image and the PAN image as a linear combination of the MS bands whose weights are estimated from the available data

Summary

Introduction

Remote sensing sensors simultaneously capture a Multispectral (MS) low resolution image along with a single-band high resolution image of the same area, referred to as Panchromatic (PAN) image.MS high-resolution images are needed by many applications, such as land use and land cover analyses or change detection. In this paper we formulate the pansharpening problem following the Bayesian framework Within this framework, we use the sensor characteristics to model the observation process as a conditional probability distribution. The observation process describes both the MS high resolution image to MS low resolution image relationship and how the PAN image is obtained from the MS high resolution one. This probability distributions provides fidelity to the observed data in the pansharpened image reconstruction process. Together with from fidelity to the data, Bayesian methods incorporate prior knowledge on the MS high resolution image in the form of prior probability distributions This probability distributions provides fidelity to the observed data in the pansharpened image reconstruction process. together with from fidelity to the data, Bayesian methods incorporate prior knowledge on the MS high resolution image in the form of prior probability distributions

Objectives

Discussion

Conclusion