Advanced Remote Sensing Technology for Sustainable Land Development in Arid Lands

Abstract

Researchers have recognized the potential use of multi-sensor remote sensing as one of the main datasources for monitoring and assessing the impact of rapid land use changes especially in fragileecosystems such as drylands. Most mapping methods have been developed for multi spectral images sincethis type of imagery currently provides the longest Earth Observing (EO) records that can be used toconstruct time series for monitoring the status of arid and semi-arid ecosystems. The spatial resolution ofolder generation EO sensors imposes, however, a problem in relation to the spatial variability ofecosystem components and status, especially in urban areas. Several techniques have been developed toimprove the detection and mapping of surface features that occur at sub-pixel resolution so that theirspectral information can be used for feature identification and extrapolated in space and time. However,ground observation is often needed to characterize and verify the land surface components making up apixel as well as to relate their spectral response to their actual physical stage (e.g., vegetation health, soilsalinity and moisture content etc.). When ground information is not available or feasible, high resolutionimagery can provide the required spatial information. This paper discusses current techniques in multisensordata analysis by illustrating their applications in urban and agricultural settings of arid lands. Keywords: multispectral, hyper spectral data, object-based classification, spectral angle mapper, aridland applications