A comparative study of NDBI, NDISI and NDII for extraction of urban impervious surface of Dehradun [Uttarakhand, India] using Landsat 8 imagery

Estimating the impervious surface is important in monitoring the spread of urban areas and human activities. This paper compares three indices, namely, Normalized Difference Impervious Index (NDII), Normalized Difference Impervious Surface Index (NDISI) and Normalized Difference Built-up Index (NDBI) for impervious surface extraction. Landsat 8 (OLI/ TIRS) imagery and LISS III data were used to extract impervious surface of Dehradun, Uttrarakhand, India. The images were acquired on November, 2015 for Landsat 8 and March 2013 for LISS III. Because of cloud free atmospheric conditions, no Atmospheric Correction is done but Dark Object Subtraction and Radiometric Correction are some of the corrections done before pre-processing. Six end members, namely, impervious surface, barren land, agricultural land, water, mountains and forest were selected for Land Use Land Cover classification. Supervised Classification (SC) of Support Vector Machine (SVM) method is used to classify impervious surface and it was observed that the Green and Thermal IR band for NDII show the maximum accuracy. User's accuracy, Producer's accuracy and Kappa cofficient are calculated and compared for all above indices.