Abstract

Satellite sensors usually provide two types of data: panchromatic and multispectral images which are characterized by their high spatial resolution and high spectral resolution respectively. In this context, the fusion techniques or pansharpening consist of merging these different aspects to obtain a fused (or pan-sharpened) image with high spatial and spectral resolutions. In this paper, a new quality assessment scheme for pan-sharpened remote sensing imagery is proposed. The methodology described extracts the segments of the images to constitute the basic elements of the measuring quality methodology. This new strategy overcomes traditional pixel-based perspectives, approaching an evaluation by human observers. The results of its application to a set of fused images show that an object-based assessment is consistent in terms of quality determination of both the spectral and spatial properties of remote sensing images.