ESA ASIMOV Project Active Fire Dataset

The ESA funded ASIMOV project is focused on designing, developing, and evaluating trustworthy Artificial Intelligence (AI) architectures for generating downscaled Essential Climate Variables (ECVs), with a specific emphasis on Active Fire detection using data from the Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard the Meteosat Second Generation (MSG) satellites. To support this goal, data from multiple sources have been compiled to create a comprehensive dataset for training deep learning-based super-resolution (SR) models aimed at enhancing wildfire detection and monitoring capabilities with SEVIRI observations. The dataset comprises 12 SEVIRI spectral bands paired with MODIS Active Fire product data, collected during fire events reported by the Greek Fire Brigade between 2014 and 2023. Additionally, topographic features such as elevation, slope, and topographic position index—derived from the GTOPO30 Digital Elevation Model—are included for each recorded fire event.由欧洲空间局（European Space Agency，ESA）资助的ASIMOV项目，致力于设计、开发与评估可信赖的人工智能（Artificial Intelligence，AI）架构，用于生成降尺度基本气候变量（Essential Climate Variables，ECVs），并特别聚焦于利用静止气象卫星第二代（Meteosat Second Generation，MSG）星载旋转增强型可见红外成像仪（Spinning Enhanced Visible and Infrared Imager，SEVIRI）的数据开展活跃火点探测。 为支撑该项目目标，研究团队整合多源数据构建了一套完备的数据集，用于训练基于深度学习的超分辨率（Super-resolution，SR）模型，以借助SEVIRI观测数据提升野火探测与监测能力。该数据集包含12个SEVIRI光谱波段数据，以及配套的中分辨率成像光谱仪（Moderate Resolution Imaging Spectroradiometer，MODIS）火点产品数据，数据采集自2014年至2023年间希腊消防部门上报的野火事件时段。此外，针对每一起记录在案的野火事件，数据集还纳入了由GTOPO30数字高程模型（Digital Elevation Model，DEM）提取的地形特征，包括高程、坡度与地形位置指数（topographic position index）。