Evaluation of spectral built-up indices for impervious surface extraction using Sentinel-2A MSI imageries: A case of Addis Ababa city, Ethiopia

Abstract

Impervious surface extraction with high accuracy is important for monitoring urban expansion to sustainably manage the land resources and save the environment. In this context, use of spectral built-up indices has been extensively explored. This study examines the performance of Built-up Area Extraction Index (BAEI), Band Ratio for Built-up Area (BRBA), Modified Built-up Index (MBI), Normalized Built-up Area Index (NBAI), New Built-up Index (NBI), Normalized Difference Built-up Index (NDBI), and Urban Index (UI) in the classification and change detection of impervious surfaces using sentinel-2A MSI imagery. All these maps were classified into built-up and non-built-up areas, and evaluated based on histogram overlap and spectral discrimination index (SDI). Simultaneously, support vector machine (SVM) algorithm was employed to classify the imageries into land-use and land-cover (LULC) classes, viz., bare land, built-up, forest, vegetation, and water bodies. The findings indicate that NBAI, NBI, and NDBI have the highest SDI values of 1.24, 1.23 and 1.54; 1.06, 1.08, and 1.23; 1.16, 1.1 and 1.26 for 2016, 2018, and 2020, respectively. However, other indices show unsatisfactory results. The LULC change between 2016 and 2020 showed that the built-up area, bare land, and water bodies have increased by 9,084.5 ha, 813 ha, and 2 ha, respectively whereas vegetation and forest areas declined by 9,279.3 ha and 620.2 ha, respectively. The classification overall accuracy was 81%, 86%, and 82% for 2016, 2018, and 2020 images, respectively thereby affirming that spectral built-up indices in urban environments can deduce impervious surface extraction quickly and accurately.