A Novel Approach to Remotely Detect Coal Mining Sites Using Hyperspectral Images

Abstract

In this work, we attempt to detect coal mining sites using a hyperspectral image in the visible and near infrared region. The hyperspectral image consisting of 17 contiguous spectral bands in the visible and near infrared spectral regions is used to determine the best combination of normalized differential spectral indices using a random forest classification approach that was trained on the ground truth binary classification map of coal mining sites. The performance metric of mean accuracy decrease highlighted the best spectral band combinations in the visible and near infrared regions that were used to detect the coal mining sites. The results were compared with methodologies available in literature, followed by statistical interpretation.