MTBF-33

MTBF-33是由科罗拉多大学博尔德分校的研究团队创建的多时相建筑轮廓数据集，涵盖美国33个县的建筑数据，时间跨度从1900年至2015年。该数据集包含超过620万个建筑轮廓及其建造年份，数据来源于公开的行政机构资源，如税务评估数据和建筑轮廓数据。创建过程中，研究团队整合和协调了这些数据，使用GeoPandas和ESRI ArcPy等工具进行数据处理。MTBF-33主要用于城市变化检测、长期土地开发分析和建筑环境研究，旨在解决历史建筑分布的量化问题，支持从遥感数据、历史地图等提取历史信息的统计学习方法。

MTBF-33 is a multi-temporal building footprint dataset developed by a research team from the University of Colorado Boulder. It covers building data across 33 counties in the United States, with a temporal range from 1900 to 2015. The dataset contains over 6.2 million building footprints along with their respective construction years, and is sourced from public administrative resources such as tax assessment data and building footprint datasets. During the creation process, the research team integrated and harmonized these datasets, and conducted data processing using tools including GeoPandas and ESRI ArcPy. MTBF-33 is primarily intended for urban change detection, long-term land development analysis, and built environment research. It aims to address the quantification of historical building distributions, and supports statistical learning approaches that extract historical information from remote sensing data, historical maps, and other similar sources.