ESA-PhiLab-Edge/WorldFloods-Phisat-2

WorldFloods Φsat-2数据集是一个模拟Φsat-2任务的数据集，用于训练基础模型。该数据集基于WorldFloods v2数据集，使用OrbitalAI Challenge中的代码模拟Φsat-2图像，并遵循PhiSatNet仓库的数据规范格式存储。数据集由非重叠的训练和验证块组成，每个块大小为512x512像素，且云量小于20%。测试集则保留了完整场景，以便更好地评估洪水检测模型。整个数据集的大小为389 GB。

The WorldFloods Φsat-2 dataset is an adaptation of the WorldFloods v2 dataset, created to simulate a Φsat-2-like dataset for training Foundation Models for this mission. Φsat-2 images were simulated using the code in the OrbitalAI Challenge and stored following the PhiSatNet repository Data Specification Format. The dataset consists of non-overlapping Train and Validation patches of 512x512 pixels with less than 20% cloud cover. The full scenes are kept for Test samples to better evaluate flood detection models. The total size of the dataset is 389 GB.