融合建筑物与POI数据的精细人口空间化研究

Fine-Scale Population Spatialization by Synthesizing Building Data and POI Data

精细的人口分布数据能刻画出行政单元内部细节的人口空间分布信息,为城市规划、灾害评估等相关研究和应用提供有效数据支撑。利用房屋建筑及高德兴趣点(point of interest,POI)数据提取建模因子,结合随机森林模型获取了武汉市2015年常住人口50 m空间化数据集。结果表明,相较于WorldPop数据集结果,所提出方法的结果在武汉市高、中、低3种不同人口密度社区单元均具有更高的拟合精度。

Fine-scale population distribution data can depict detailed spatial distribution information of population in administrative unit,which can provide effective data support for urban planning and disaster assessment. The modeling factors extracted from the building data and point of interest(POI)data are used with random forest model to obtain the spatial data set of permanent population with 50 m resolution in Wuhan in 2015. Compared with the result of WorldPop dataset,the result of the proposed method has higher fitting accuracy in community units of Wuhan with high,medium or low population densities.