Schultz Michael, Li Hao, Wu Zhaoyan, Wiell Daniel, Auer Michael, Alexander Zipf
Geoinformatics of University of Heidelberg, Heidelberg, Germany.
Department of Geography, National University of Singapore, Singapore, Singapore.
Sci Data. 2025 May 6;12(1):750. doi: 10.1038/s41597-025-04703-8.
Our map represents the first successful large-area fusion of OpenStreetMap and Copernicus data at a spatial resolution of 10 m or finer and can be applied globally. We addressed varying label noise and feature space quality, utilizing artificial intelligence and advanced computing. Our method relies solely on openly available data streams and methods, eliminating training data acquisition or the need for additional expert knowledge for such purpose. We extracted land use labels from OpenStreetMap and remote sensing data to create a contiguous land use map of the European Union as of March 2020. OpenStreetMap tags were translated into land use labels, directly mapping 61.8% of the Union's area. These labels served as training data for a classification model, predicting land use in remaining areas. Country-specific deep learning convolutional neural networks and Sentinel-2 feature space composites of 2020 at 10 m resolution were employed. The overall map accuracy is 89%, with class-specific accuracies ranging from 77% to 99%. The data set is available for download from https://doi.org/10.11588/data/IUTCDN and visualization at https://osmlanduse.org .
我们的地图代表了首次成功地以10米或更精细的空间分辨率对OpenStreetMap和哥白尼数据进行大面积融合，并且可以在全球范围内应用。我们利用人工智能和先进计算技术，解决了不同的标签噪声和特征空间质量问题。我们的方法仅依赖于公开可用的数据流和方法，无需为此目的获取训练数据或额外的专家知识。我们从OpenStreetMap和遥感数据中提取土地利用标签，以创建截至2020年3月的欧盟连续土地利用地图。OpenStreetMap标签被转换为土地利用标签，直接覆盖了欧盟61.8%的面积。这些标签用作分类模型的训练数据，用于预测其余地区的土地利用情况。采用了特定国家的深度学习卷积神经网络和2020年10米分辨率的哨兵-2特征空间合成数据。地图的总体准确率为89%，特定类别的准确率在77%至99%之间。该数据集可从https://doi.org/10.11588/data/IUTCDN下载，并可在https://osmlanduse.org上进行可视化。
