Antarctic land-cover classification using IKONOS and Hyperion data at Terra Nova Bay

Abstract

Low or medium spatial resolution satellite images are used for environmental monitoring in remote, extremely cold areas such as Antarctica. However, they cannot provide detailed spatial and spectral information over Antarctic areas. For obtaining this information, we propose an Antarctic land-cover classification method using IKONOS and Hyperion data over Terra Nova Bay, Antarctica. High spatial resolution IKONOS imagery enabled the detection of detailed, accurate boundaries between areas of snow, ice, vegetation, water and rock. Rock types were classified in detail using high spectral resolution Hyperion imagery. Because Antarctic land-cover types exhibit unusual spectral features, a step-by-step classification using various data sets and methods was applied. Five soil types and two rock types were identified by applying spectral linear unmixing to Hyperion data. Our land-cover classification map shows small mossy areas near emperor penguin or skua habitat. Lakes, which are an important factor in environmental monitoring, are also clearly shown. The detailed rock and soil classification map revealed that biotite gneiss and clay soils are dominant. This Antarctic land-cover map is useful for determining a suitable site for an environment-friendly Antarctic station. Among the various land-cover types, the lakes and the vegetation, rock and soil areas are considered to be important environmental factors related to Antarctic living organisms.

Acknowledgements

This research was supported by a grant from a Strategic Research Project (Development of Site Investigation and Monitoring System for Extreme Cold Region) funded by the Korea Institute of Construction Technology.