Context-adaptive pansharpening algorithm for high-resolution satellite imagery

Abstract

Pansharpening algorithms are important methods for overcoming the technical limitations of satellite sensors. However, most approaches to pansharpening have tended either to distort the spectral characteristics of the original multispectral image or reduce the visual sharpness of the panchromatic image. In this paper, we propose a pansharpening algorithm that uses both a global and a local context-adaptive parameter based on component substitution. The purpose of this algorithm is to produce fused images with spectral information similar to that of multispectral data while preserving spatial sharpness better than other algorithms such as the Gram–Schmidt algorithm, Additive Wavelet Luminance Proportional algorithm, and algorithms developed in our previous work. The proposed parameter is calculated using the spatial and spectral characteristics of an image. It is derived from a spatial correlation between each multispectral band and adjusted-intensity image based on Laplacian filtering and statistical ratios, and the parameter is subsequently adaptively adjusted using image entropy to optimize the fusion quality in terms of the sensor and image characteristics. An experiment using IKONOS-2, QuickBird-2, and Geoeye-1 imagery demonstrated that a global context-adaptive parameter model is effective for spatial enhancement and that a local context-adaptive model is useful for image visualization and spectral information preservation.