巴基斯坦2023年众(开)源建筑融合数据集OA
A dataset of multi-source crowdsourced/open building data fusion in Pakistan in 2023
建筑空间数据是灾害暴露评估、建成环境分析和应急救援的重要基础数据.针对现有开源建筑数据在时间基准不统一、几何表达差异大、属性完整性不足以及跨源冲突明显等问题,本研究构建了巴基斯坦 2023年多源开源建筑融合数据集.本数据以 Google Open Buildings、Microsoft Global ML Building Footprints 和OpenStreetMap 建筑数据为基础,辅以 2023年建筑存在支持强度栅格、建筑高度栅格和道路网络信息,通过"预处理—时效性验证—匹配融合—道路约束平差—质量控制"流程生产.数据集以 GeoPackage 格式组织为统一建筑主数据,以建筑面要素为基本单元,记录目标几何轮廓、来源标记、时效状态、匹配关系、平差参数及建筑类型、高度等补充属性.质量评估结果显示,建筑间总重叠面积降低97.83%,建筑—建筑冲突对解决率达86.99%,道路—建筑冲突对解决率达84.93%;平均交并比为 0.826,全局面积损耗率为 2.63%,建筑平均位移量控制在 0.32 m.本数据集具有统一时间基准下的建筑存量筛查、空间冲突消解和结果可追溯表达等特点,可为巴基斯坦的灾害评估、暴露建模、建成区分析及应急救援等研究与应用提供可靠基础数据支撑.
Building spatial data are fundamental to disaster exposure assessment,built environment analysis,and emergency response.To address the critical limitations of inconsistent temporal baselines,significant geometric discrepancies,incomplete attribute information,and prominent cross-source conflicts in existing open building datasets,this study presents the dataset of multi-source crowdsourced/open building data fusion in Pakistan in 2023.The dataset is constructed based on Google Open Buildings,Microsoft Global ML Building Footprints,and OpenStreetMap building data,supplemented by 2023 building-presence probability rasters,building-height rasters,and road network information.It is produced through a standardized workflow consisting of preprocessing,temporal validity assessment,matching and fusion,road-constrained adjustment,and quality control.Organized as a unified building master dataset in GeoPackage format,the dataset adopts building polygon features as the basic spatial unit.Each feature records comprehensive information,including target geometric footprints,source labels,temporal validity status,matching relationships,adjustment parameters,and supplementary attributes such as building type and height.Quality assessment results demonstrate that the total overlapping area between buildings was reduced by 97.83%;the resolution rate of building-building conflict pairs reached 86.99%,and the resolution rate of road-building conflict pairs reached 84.93%.The mean Intersection over Union(Mean IoU)is 0.826,the global area loss rate is 2.63%,and the average building displacement is controlled at 0.32 m.This dataset features building stock screening under a unified temporal baseline,systematic spatial conflict resolution,and fully traceable representation of fusion results.It provides reliable fundamental data support for research and applications,including disaster risk assessment,exposure modeling,built-up area analysis,and emergency response in Pakistan.
陈诺;周晓光;王岚;侯东阳;Karamat Ali;彭舒;林尚纬
中南大学地球科学与信息物理学院,长沙 410000中南大学地球科学与信息物理学院,长沙 410000中南大学地球科学与信息物理学院,长沙 410000中南大学地球科学与信息物理学院,长沙 410000喀喇昆仑国际大学,巴基斯坦吉尔吉特 15100国家基础地理信息中心,北京 100830国家基础地理信息中心,北京 100830
开源建筑数据建筑融合时效性验证空间冲突消解巴基斯坦
open-source building databuilding data fusiontimeliness validationspatial conflict resolutionPakistan
《中国科学数据(中英文网络版)》 2026 (3)
133-151,19
国家重点研发计划(2023YFE0208000) National Key Research and Development Program of China(2023YFE0208000)
评论