MULTITEMPORAL ANALYSIS OF LANDSAT IMAGES IN FORMER TIN ALLUVIAL MINING AREA, BANGKA ISLAND

• Main Author: Syamsul Bahri (NIM : 12114016), Muslim
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/29408
• Institution: Institut Teknologi Bandung

This study was conducted to perform multitemporal analysis of Landsat images on the mining area of tin alluvial deposit. Land cover classification maps were produced using a supervised classification method, which was Maximum Likelihood Classification (MLC). In performing MLC, region of interest (ROIs) must be made first as training data. ROIs were made with help of two indices: Normalized Difference Vegetation Index (NDVI) to extract vegetation information; and Modified Normalized Difference Water Index (MNDWI) to extract water-body features. Spectral Angle Mapper (SAM) method was then applied to map the area with the similar reflectance as the tailing samples. Reflectance data of the tailing samples retrieved from ground-check showed that the tailing consisted mostly of quartz and kaolinite/smectite clay. Results of land cover classification and tailing-covered area mapping showed that the mining activity in study site has been increasing, drastically from 2005 to 2008. Spectral analysis of the reflectance data from tailing samples showed that tailing is mostly consisted of quartz, kaolinite, and smectite.