Delineating urban, suburban and rural areas using Landsat and DMSP-OLS night-time images

摘要

The number and size of urban settlements are increasing in all the continents of the world at a rapid pace. Urban sprawl is associated not only with changes in landcover and area, but also ecological, climate and social transformations. Mapping the growth and spread of urban areas is important. Remote sensing has long been used to map human settlements. Today the availability of a large number of satellites and sensors, determining the appropriate image to map urban area is a research area itself. This study compares two satellite images: Landsat Enhanced Thematic Mapper data and Defence Meteorological Satellite Program, Operational Linescan System image to map the urban footprint of the city of Hyderabad, India. Landsat ETM data is captured during the daytime and gives spectral reflectance values while the DMSP-OLS data captures artificial lights from human settlements at night and produces brightness information. The results show an accuracy of more than 90% in the classification and delineation of urban, suburban and rural landcover types. This study shows that in addition to spectral reflectance captured by satellites from different features on the earth surface during the daytime, differences in the degree of brightness of the lights emitted from urban areas at night is also an effective indicator in delineating landcover types.