Spectral Unmixing Applied to Desert Soils for the Detection of Sub- Pixel Disturbances.

摘要

Desert areas cover approximately one-fifth of the Earth, making it important to understand how disturbance affects arid regions on a spectral level. Remote sensing technology can be used to detect and characterize surface disturbance both literally (visually) and non-literally (analytically). Non- literal approaches may even allow detection of anthropogenic-related surface disturbances that are not visible in individual images or color composites. This is achievable through identification of differences in spectral reflectance among like soil components, both chemical and biological. Previous research suggests that surface disturbances cause alteration of soil properties, making it feasible to detect variation in reflectance signatures. This research supports that assumption and has determined that disturbance-related changes do have unique spectral characteristics in hyperspectral imagery that are detectable, even at the sub-pixel level and using endmembers from geographically different yet geologically similar regions.