Assessment of open-pit captive limestone mining areas using sentinel-2 imagery with spectral indices and machine learning algorithms.

This article focuses on mapping and assessing active captive limestone mining areas in the Yerraguntla cement industrial region, Andhra Pradesh, India, using remote sensing data and machine learning algorithms for the financial year 2019. Utilizing high-resolution Sentinel-2A satellite imagery, the study applied spectral index (Normalized Difference Vegetation Index, NDVI), unsupervised (ISODATA), and supervised machine learning classifiers—K-Nearest Neighbors (KNN) and Random Forest (RF)—to estimate mining extents and compare them with industrial field survey data. The RF classifier achieved the highest overall accuracy (95.79%) and kappa coefficient (0.957) but underestimated the mining area (379.57 ha) relative to industrial data (487.10 ha), while KNN also performed well (94.78% accuracy, 417.47 ha area). The NDVI method provided a closer area estimate (469.92 ha) but with lower accuracy (79.84%), and ISODATA showed the poorest performance (64.16% accuracy). The study demonstrates the utility of machine learning and remote sensing for environmentally sustainable mine monitoring and land use/cover classification in limestone mining regions.