A CRITICAL REVIEW OF QUALITY ASSESSMENT PROTOCOLS IN PAN-SHARPENING

Abstract. Although vast amounts of pan-sharpening methods have been proposed to date, there has been relatively little published on the topic of qualitative and quantitative assessment of the pan-sharpened multispectral (MS) data. Since a high resolution reference MS image is not available, qualitative and quantitative assessments of the spatially enhanced MS image are a much more challenging task in pan-sharpening. Thomas et al. conducted a critical survey of conventional pan-sharpening methods considering remote sensing physics. In this paper, we study the effects of physical constraints of satellite sensors on conventional quality assessment protocols. The novelty of this work is in investigating the effect of physical constraints on the performance of quality assessment protocols. The most popular protocols, which are analysed here, are quality not requiring a reference (QNR), Wald’s protocol, Zhou’s protocol, Khan’s protocol and Ghassemi’s protocol. In doing so, the strengths and weaknesses of such protocols regarding the physical limitations on satellite sensors are discussed.