Indian Journal of Science and Technology

Abstract

Objective: Mapping impervious surfaces using moderate resolution satellite images is a useful technique for supporting different fields including monitoring and evaluation, planning, statistical analysis and reporting and policy development. Due to the negative impact of impervious surfaces over urban climate, the development of new techniques using spectral indices is a key parameter to extract built up areas with high accuracy. Method: In this study we propose a new concept capitalizing the existing relationships between urban heat island effects and built up areas to produce a modified index, benefiting from the high reflectivity of thermal bands, the mean-infrared and near-infrared band. In addition, spatial enhancement of multispectral data is also used to improve the accuracy of the proposed index, our approach uses spectral reduction of dimensions in order to produce thematic indices (as an input data) instead of continuous images. Finding: Final result showed a high accuracy of built up areas compared to other spectral indices like normalized difference built-up index - NDBI and index -based built-up index - IBI, more than 10% compared to classic NDBI and 6% compared to modified index (IBI).
Keywords: Landsat OLI, PCA, Remote Sensing, Spatial Enhancement, Spectral Index, Urban Heat Island Effect