3D-GloBFP：全球建筑物高度数据集

全球建筑物高度数据集（3D-GloBFP）是首个在单体建筑轮廓级别上实现全球覆盖的建筑物高度数据集，提供了2020年全球约20亿建筑的三维结构信息。该数据集融合多源地球观测（EO）数据，并基于极端梯度提升模型（XGBoost）构建而成，其准确性性与精度已在全球33个子区域中得到验证，决定系数（R²）介于 0.66 至 0.96 之间，均方根误差（RMSE）为 1.9 米至 14.6 米。3D-GloBFP 为深入理解人类与建成环境之间的复杂关系提供了坚实的数据基础，具有在气候建模、能源消耗评估、城市形态分析及社会经济研究等多个领域的广泛应用潜力和重要价值。目前，该数据集描述论文已在地球科学数据权威期刊《Earth System Science Data》上发表（链接：https://essd.copernicus.org/articles/16/5357/2024/）。

The Global Building Height Dataset (3D-GloBFP) is the first building height dataset with global coverage at the individual building footprint level, providing three-dimensional structural information for approximately 2 billion buildings worldwide in 2020. This dataset is developed by fusing multi-source Earth Observation (EO) data and leveraging the Extreme Gradient Boosting (XGBoost) model. Its accuracy and precision have been validated across 33 global sub-regions, with the coefficient of determination (R²) ranging from 0.66 to 0.96 and the root mean square error (RMSE) between 1.9 meters and 14.6 meters. 3D-GloBFP provides a solid data foundation for in-depth understanding of the complex relationship between humans and the built environment, and has broad application potential and significant value in multiple fields including climate modeling, energy consumption assessment, urban morphology analysis, and socio-economic research. Currently, the paper describing this dataset has been published in the authoritative geoscience data journal *Earth System Science Data* (link: https://essd.copernicus.org/articles/16/5357/2024/).

AI搜集汇总

背景与挑战

背景概述

3D-GloBFP是全球首个在单体建筑轮廓级别覆盖全球的建筑物高度数据集，提供2020年约20亿建筑的三维信息，基于多源地球观测数据和XGBoost模型构建，验证精度较高（R²为0.66-0.96，RMSE为1.9-14.6米）。该数据集支持气候建模、能源评估等应用，数据开放获取，大小为152.22 GB。