Mining activities significantly impact the environment, necessitating effective, continuous monitoring. Traditional surface monitoring methods are often costly and labor-intensive. This study proposes an automated workflow using the SpatioTemporal Asset Catalog (STAC) and Sentinel-2 satellite imagery to monitor mining surface changes. By calculating the Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), and Modified Bare Soil Index (MBI), the workflow identifies land cover changes within mining concessions. The system was Implemented in Python environment using libraries such as PySTAC, PySTAC Client, Xarray, Rioxarray, Geopandas, Dask, and Numpy. The mining surface change was analyzed using the regression line gradient of each spectral index. Results show active mining sites exhibit an NDVI slope lower than -1, indicating rapid conversion of vegetation to non-vegetative land due to land clearing activities. Conversely, the positive NDWI trend indicates increased water coverage from land excavation, while the MBI trend is the weakest, suggesting limited sensitivity to surface changes in mining areas. To evaluate the accuracy of the results, manual verification was conducted. The analysis revealed that 3 out of 25 mining concessions were incorrectly classified, resulting in an overall accuracy of 88%.
Mulkal et al. (Mon,) studied this question.