Abstract

RADARSAT-2 data in the ultra-fine and fine quad polarization mode was evaluated to identify morpho-structural, periglacial, and other geomorphological features within the Antarctic Peninsula region. The study was carried out within Byers Peninsula, the largest ice-free area on the South Shetland Islands. A methodology was developed to integrate and compare data from different sources to optimize the selection of criteria and techniques to obtain information from RADARSAT-2 relevant to the geomorphology of the study area. Ultra-fine data were focused on the extraction of linear features and fine structures; whereas, fine quad polarized data were implemented to identify scattering mechanisms for different surface covers. It was possible to extract a greater number of lineaments and structures with the ultra-fine data. A supervised classification was carried out using the fine quad polarized data, where a careful selection of training and validation sites was needed. The overall classification accuracy was 75% with a Kappa coefficient of 0.71. Surface cover features with low entropy scattering mechanisms were well identified with a maximum of 86% accuracy. The accuracy for higher entropy volume and multiple scattering mechanisms were only slightly lower and are considered especially important for the type of geomorphological features studied on Byers Peninsula.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call