Improvement of effective spatial resolution of thermal infrared data for urban landuse classification

Abstract

Remotely sensed thermal infrared (TIR) data contains information on emitted radiation that is complementary to that in the visible (VIS) and near infrared (NIR) spectrum. Despite the unique radiometric properties, the application of TIR data has been rather restricted, especially in urban areas that require higher spatial resolutions for accurate classification, mainly due to its coarse spatial resolution. This paper presents the methods adopted to improve the effective spatial resolution of TIR data and use its complementary information content for improved urban landuse classification. In this study spatial filtering and data fusion are applied to improve the effective spatial resolution of Landsat-5 TM band 6 data of Mumbai, a metropolitan city in western India. Supervised classification was performed, without the input of TIR data and with the input of improved spatial resolution TIR data and the results are compared.