Signal-dependent noise modelling and estimation of new-generation imaging spectrometers

The application of space technology to archaeological research has been paid great attention worldwide, mainly because the current availability of very high resolution (VHR) satellite imagery, such as, IKONOS (1999) and QuickBird (2001), provide valuable data for searching large areas to find potential archaeological sites. Data from VHR satellite can be very useful for the identification, management and documentation of archaeological resources. Archaeological investigation based on the use of VHR satellite images may take benefits from the integration and synergic use of both panchromatic and multispectral data. This can be achieved by using pansharpening techniques, which allow multispectral and panchromatic images to be merged. The two basic frameworks of pansharpening techniques are Component Substitution (CS), such as Intensity-Hue-Saturation (IHS) Gram-Schmidt (GS), and multiresolution analysis (MRA), such as wavelets and Laplacian pyramids (LP). In this paper, both Gram-Schmidt and Laplacian pyramids with context adaptive (CA) detail injection models were used. QB images were processed for a relevant archaeological area in Southern Italy, the ancient Siris-Heraclea, a very significant test area because it is characterized by the presence of both surface and subsurface ancient remains. Outcomes of different pansharpening techniques have been qualitatively evaluated for both surface and subsurface remains. The visual inspection clearly suggests that the quantitative evaluation of the fusion performance for archaeological applications is a critical issue, and "ad hoc" local (i.e. context-adaptive) indices need to be developed.