Authors: Jeffrey Colby*, Appalachian State University, Michael Bishop, Texas A&M University
Topics: Remote Sensing, Mountain Environments
Keywords: remote sensing, mountain environments, topographic correction, multi-spectral imagery
Session Type: Paper
Start / End Time: 3:05 PM / 4:45 PM
Room: Buchanan, Marriott, Mezzanine Level
Presentation File: No File Uploaded
According to the Worldatlas, mountains cover 24% of the world’s landmass. The remote locations of mountain regions and the urgency for studying environmental change, natural hazards, and water resource issues dictate the need for the acquisition and interpretation of multi-spectral imagery. Effective analysis of multi-spectral imagery in areas of complex topography, however, is compromised by multi-scale topographic effects that govern reflectance. As research to address this problem has evolved in the last 40 years, it has become evident that components of the radiant transfer cascade must be accounted for including orbital and solar geometry, atmospheric constituents, topography, biophysical properties, land cover structure and sensor characteristics. The term used to describe these factors for topographic correction is called anisotropic reflectance correction (ARC).
A review of the literature indicates that topographic correction methods initiated in the late 1970’s have focused on spectral feature extraction, empirical, semi-empirical, and radiation transfer modeling approaches. In the last decade, over a dozen articles have been published comparing empirical and semi-empirical algorithms. Results from these studies are inconclusive overall and lack a formalized structure for treatment of topographic complexity and application of evaluation methods. Our meta-analysis of results indicate that we may have reached a plateau in terms of applying our understanding of topographic factors and making progress in developing new parameterization schemes for topographic correction. Moreover, the lack of progress in developing more comprehensive ARC approaches suggests that we may need to re-evaluate fundamental assumptions upon which current empirical and semi-empirical approaches are based.