Detection of Coal Quarry and Coal Dump Regions Using the Presence of Mine Water Bodies from Landsat 8 OLI/TIRS Images

Jit Mukherjee, Jayanta Mukherjee, Debashish Chakravarty

Abstract: Surface mining has major environmental, social, and economical adversities, which makes it an active area of research in remote sensing. Surface coal mining has additional adversities of coal seam fires. Thus, the detection, classification, and monitoring of such regions have various research challenges. The surface coal mining land classes cover smaller areas compared to mid-resolution satellite images making them challenging to detect. Coal quarry and coal dump regions are such kinds of smaller land classes. They can be detected as a single land class as discussed in the literature. However, these land classes are observed to be difficult to detect separately as they follow near similar spectral characteristics. Hence, this paper proposes a novel technique to separate these regions using the presence of water bodies. Coal dump regions do not have water bodies, whereas some coal quarry regions may have water bodies. Such quarry regions are detected at first and further, they are used to train an unsupervised single class support vector machine (SVM). This model is used to detect the coal dump regions by detecting the outliers. The proposed technique provides average precision and recall for coal quarry, and coal dump regions as $$[84.88\%,61.44\%]$$ , and $$[70.91\%,52.79\%]$$ , respectively over the seasons.