MapInWild

MapInWild数据集是由德国联邦国防军大学空间技术和空间应用研究所航空航天工程系的研究人员创建，旨在通过遥感技术解决自然环境中的荒野地图绘制问题。该数据集包含8144张1920×1920像素的图像，涵盖多种传感器数据，如Sentinel-1和Sentinel-2，以及ESA WorldCover和VIIRS等。数据集通过Google Earth Engine平台进行管理和查询，确保数据的时空一致性和可重复性。MapInWild数据集不仅服务于环境遥感领域，还扩展到机器学习研究，特别是在弱监督学习方面，为理解和保护自然环境提供了新的视角和工具。

The MapInWild dataset was created by researchers from the Department of Aerospace Engineering, Institute of Space Technology and Space Applications, the University of the Federal Armed Forces Germany. It was developed to address the challenge of wilderness mapping in natural environments using remote sensing technologies. This dataset includes 8,144 images with a resolution of 1920×1920 pixels, encompassing multi-source sensor data such as Sentinel-1, Sentinel-2, ESA WorldCover, and VIIRS, among others. The dataset is managed and queried through the Google Earth Engine platform, which ensures the spatiotemporal consistency and reproducibility of the data. Beyond serving the field of environmental remote sensing, the MapInWild dataset also extends to machine learning research, especially in the area of weakly supervised learning, providing novel perspectives and tools for understanding and conserving natural environments.