基于NPP/VIIRS夜间灯光数据和土地利用数据的人口分布图绘制——以大连金普新区为例  

Population Distribution Map Drawing Based on NPP/VIIRS Nighttime Light Remote Sensing Images and Land Use Data:A Case Study in Jinpu New Area

作　　者：李欣欣 王利[1] 何飞[1] LI Xinxin;WANG Li;HE Fei(College of Urban and Environment,Liaoning Normal University,Dalian,Liaoning 116029,China)

机构地区：[1]辽宁师范大学城市与环境学院,辽宁大连116029

出　　处：《遥感信息》2018年第4期35-41,共7页Remote Sensing Information

基　　金：国家科技支撑计划(2008BAH31B06);部省共建人文社会科学重点研究基地项目(14JJD790038)

摘　　要：针对现有人口分布图绘制不够精确,分布状况偏离实际的问题,提出一种利用VIIRS DNB数据对人口进行空间化的方法。以大连市金普新区为例,采用土地利用数据、新一代NPP/VIIRS夜间灯光数据分别建立0.5km×0.5km格网人口预测模型,根据构建的非户籍常住人口分配系数,对建立的模型进行街道尺度的精度评价。结果表明,基于土地利用数据和灯光数据建立的预测模型整体误差分别为0.279和0.142,表明NPP/VIIRS夜间灯光数据能够更为准确地模拟人口,最后根据基于灯光数据建立的人口模型预测得到的人口数,依据"无居住用地则无人口原则"进行人口空间化,绘制出较为精准的人口分布图。该研究探索的技术方法和模型也可以为其他区域相关研究提供参考。In view of the problems that the existing population map is not accurate enough,and the distribution situation deviates from the fact,this paper proposed a method of spatializing the population using VIIRS DNB data and land use data.Taking the Jinpu New Area of Dalian city as an example,using the land use data and the new generation NPP/VIIRS night light data,0.5 km×0.5 km grid population forecasting model was established respectively.According to the build of household resident population distribution coefficient,it evaluated the precision of the street scale for the model.The results showed that:based on the data of land use and lighting data,the overall error of the established forecast model was 0.279 and 0.142,respectively,indicating that the NPP/VIIRS night light data can be more accurate in population simulation.Finally,based on light data,it established a model to predict the population,and according to the principle of“no residential land,no population”,it mapped the more accurate population map.This study explored the technical methods and models and can also provide a reference for other related research areas.

关 键 词：人口空间化 土地利用数据 NPP/VIIRS 人口分配系数 金普新区 

分 类 号：TP7[自动化与计算机技术—检测技术与自动化装置] K901.3[自动化与计算机技术—控制科学与工程]