基于多源信息的高空间分辨率人口分布研究

(1.西南石油大学 土木工程与测绘学院,四川 成都 610500; 2.西南石油大学测绘遥感地理信息防灾应急研究中心,四川 成都 610500)

应急救援; 主成分分析; 人口分布; POI; 建筑物

Research on Population Distribution with High Spatial Resolution Based on Multi-source Information
LIU Zhicheng1,XIAO Dongsheng1,2

(1.School of Civil Engineering and Surveying and Mapping,Southwest Petroleum University,Chengdu 610500,Sichuan,China)(2.Disaster Prevention and Emergency Research Center for Surveying,Mapping and Remote Sensing,Southwest Petroleum University,Chengdu 61050

emergency and rescue; principal component analysis; population distribution; POI; buildings

备注

针对已有的人口空间化研究多采用静态数据、时空分辨率较低、在应急救援等方面实用性不高的问题,提出了一种使用高时空分辨率数据,结合城市圈层结构理论和主成分分析法的建筑物尺度人口估算方法。以成都市为例,利用腾讯位置大数据,通过计算不同城市圈层的定位率,得到了成都市不同时段1 km×1 km的人口分布数据。在此基础上,以基于建筑物中心点的泰森多边形为人口分配基本单元,结合宜出行热力数据和POI数据,分别计算其对人口分布的贡献值并赋予计算权值,得到了成都市青羊区建筑物尺度人口分布数据。街道尺度统计数据回归分析的决定系数R2为0.926 4,总体精度较高,模拟人口分布符合实际情况。
The current spatialized data of population are normally static and of low resolution in time and space,and not so practical for emergency and rescue.To solve this problem,according to the theory of urban circle structure and principal component analysis,we proposed a method of estimating population on building-scale using high resolution data in time and space.Taking Chengdu City as an example,we calculated the location rate of different urban circles using the Tencent Location Big Data,and obtained the population distribution in the grid of 1 km×1 km of Chengdu in different periods of a day.Further,we used the Tyson Polygon based on the center of the building as the basic unit of population distribution,and calculated and weighted the contribution of the data from Tencent Easygo and the POI data from Baidu Map to the population distribution.In this way we obtained the distribution of population on building-scale in Qingyang District of Chengdu.R2,the coefficient of determination for the regression analysis of the street-scale statistics,is 0.926 4.Our results are more accurate and conform to the actual distribution of population in this area.