Nighttime Fire, Volcano, Gas Flare Detections from Ensemble Learning on NASA Black Marble

This dataset consists of ensemble machine learning detections of nighttime fires, gas flares, and volcanic erruptions from NASA Black Marble M-bands and DNB. The model is learned from unsuperivsed detections of anomalies followed by iterative training of classifiers on the detections.本数据集依托美国国家航空航天局（NASA）黑大理石（Black Marble）产品的M波段与昼夜波段（Day/Night Band，DNB）数据，涵盖针对夜间火灾、燃气火炬与火山喷发的集成机器学习检测结果。该模型首先以无监督异常检测得到的结果为训练基础，随后基于这些检测结果对分类器开展迭代训练。