地理空间分析综合应用

基于夜间灯光与LUC数据的川渝地区人口空间化研究

  • 胡云锋 , 1, * ,
  • 赵冠华 1, 2 ,
  • 张千力 1, 2
展开
  • 1. 中国科学院地理科学与资源研究所,北京 100101
  • 2. 中国科学院大学,北京 100049

*作者简介:胡云锋(1974-),男,江西赣州人,博士,副研究员,主要从事遥感监测与区域可持续发展评价研究。E-mail:

收稿日期: 2017-04-30

  要求修回日期: 2017-08-25

  网络出版日期: 2018-01-20

基金资助

国家重点研发计划项目(2016YFB0501502、2016YFC0503701);高分专项(00-Y30B14-9001-14/16)

Spatial Distribution of Population Data Based on Nighttime Light and LUC Data in the Sichuan-Chongqing Region

  • HU Yunfeng , 1, * ,
  • ZHAO Guanhua 1, 2 ,
  • Zhang QianLi 1, 2
Expand
  • 1. Institute of Geographic Sciences and Natural Resources, Chinese Academy of Sciences, Beijing 100101, China
  • 2. University of Chinese Academy of Sciences, Beijing 100049, China
*Corresponding author: HU Yunfeng, E-mail:

Received date: 2017-04-30

  Request revised date: 2017-08-25

  Online published: 2018-01-20

Supported by

National Key Research and Development Program of China, No.2016YFB0501502, 2016YFC0503701;Key Project of High Resolution Earth Observation System, No.00-Y30B14-9001-14/16.

Copyright

《地球信息科学学报》编辑部 所有

摘要

高精度的人口空间分布数据是开展小尺度人口活动变化规律研究的关键数据。夜间遥感影像对于反映人类社会活动具有独特的能力,因而被广泛的应用于社会经济领域的空间数据挖掘。本研究以DMSP/OLS夜间灯光数据、NPP/VIIRS夜间灯光数据、常住人口统计数据、土地利用数据为数据源,在县级尺度上建立逐步回归模型,构建川渝地区人口空间分布数据;并随机抽取研究区内500个乡镇,以常住人口统计数据为真实数据,对人口空间化结果进行精度检验。结果表明:① 2种夜间灯光数据与人口均有较高的相关性,相关系数均在0.76以上,NPP/VIIRS夜间灯光数据与人口的相关性要高于DMSP/OLS,且受拟合模型的影响不大。② 与人口相关性较高的土地利用类型有多种,耕地、林地对人口空间分布也有影响,在建模时不应只考虑建成区。③ 在2种夜间灯光数据分别与土地利用与土地覆被数据(Land Use/ Land Cover, LUC)结合时,使用DMSP/OLS夜间灯光数据和NPP/VIIRS夜间灯光数据得到的逐步回归模型的复相关系数R2分别为0.796、0.817,模型拟合率较高,而基于NPP/VIIRS夜间灯光数据得到的人口空间化结果分辨率(500 m)相比DMSP/OLS提高了一倍(1 km),中心城区与周边城区人口密度变化更加自然,更符合实际人口分布情况。④ 与LUC数据结合时,使用NPP/VIIRS夜间灯光数据得到的人口空间化结果精度要高于DMSP/OLS夜间灯光数据,表明NPP/VIIRS夜间灯光数据相比DMSP/OLS更适用于人口数据空间化研究。

本文引用格式

胡云锋 , 赵冠华 , 张千力 . 基于夜间灯光与LUC数据的川渝地区人口空间化研究[J]. 地球信息科学学报, 2018 , 20(1) : 68 -78 . DOI: 10.12082/dqxxkx.2018.170224

Abstract

Spatial distribution data of the population with high-precision is the important data to study the law of the variation of population activity at small-scale. Remote sensing images of nighttime light have the unique ability of reflecting human social activities. Thus, they were widely used in spatial data mining of the socio-economic field. In this study, DMSP/OLS nighttime light data, NPP/VIIRS nighttime light data, resident population data and land use data were used as data sources. Then, we used these data to build the stepwise regression at county scale. The spatial distribution data of population in Sichuan and Chongqing area were established based on the stepwise regression model. Finally, we took the resident demographic data of the randomly selected 500 townships as the practical data to assess the accuracy of spatial distribution data of the population. The analysis shows that: (1) both of the two nighttime light data have high correlation with the population. The correlation coefficients are both above 0.76. The correlation of NPP / VIIRS night light data and population is higher than DMSP / OLS. The fitting model does not change the results. (2) There are many types of land use that are highly relevant to population. Farmland and woodland can also affect the spatial distribution of population. Thus, built area should not be considered as the only type of land use for building the population distribution model. (3) When the two nighttime lights were combined with LUC( Land Use/ Land Cover), the complex correlation coefficient (R2) of the stepwise regression model using DMSP / OLS nighttime light data and NPP / VIIRS night light data is 0.796 and 0.817 respectively, and the model fitting rate is higher. Compared with the results based on DMSP/OLS (1 km), the spatial resolution of population based on NPP/VIIRS nighttime light data increases to 500 m. The change of population density is more natural from the central city to the surrounding urban area, and the population distribution is more real. (4) When combined with LUC data, the results obtained with NPP / VIIRS nighttime light data were more accurate than DMSP / OLS nighttime light data, indicating that NPP / VIIRS nighttime light data is more suitable for the research of spatial distribution of population than DMSP / OLS.

1 引言

人口问题是影响当今世界可持续发展的重大问题之一[1]。人口的增长加大了全球资源和环境的承载压力[2],资源、环境、人口之间的矛盾随之也变得十分突出[3]。掌握人口信息,研究人口空间分布及其变化可以为区域可持续发展研究、规划等工作提供科学支撑。人口调查是当前各国实现人口信息统计和分析的主要渠道,具体方式包括抽样调查和全体普查2种形式。虽然人口调查和统计有严谨的统计学理论和方法作为支撑,具有权威、系统、规范等优势[4],但仍存在时间分辨率低、更新周期长、精度低、不利于可视化和空间分析操作等问题,难以满足人口分布规律研究[5]。实时、可靠的人口空间分布信息对于研究和解释人类对社会、经济和环境的影响有重要作用[2,6-7]
人口空间分布信息可以由人口数据空间化或人口密度格网化技术获取得到。近年来,人口空间化研究发展较快,其中夜间灯光作为综合指示因子来进行社会经济数据空间化模拟得到越来越多的应用[8,9,10,11]。早期研究大多使用美国国防气象卫星(Defense Meteorological Satellite Program, DMSP)搭载的业务线扫描系统传感器(Operational Line Scan System, OLS)获取的夜间灯光数据,由于该数据空间分辨率较低(约为1 km),因此在大、中尺度(国家、省、州)上的相关研究较为适用[12,13,14],小尺度的估算适用性较低。2011年10月,美国新一代极轨运行环境卫星系统预备项目卫星(National Polar-orbiting Operational Environmental Satellite System Preparatory Project,NPP)发射成功,NPP携带的可见光红外线成像辐射仪(Visible Infrared Imaging Radiometer Suite,VIIRS)可得到接收22个波段的影像数据,其中白天/夜间波段(Day/Night Band, DNB)能够识别微弱灯光源。相较于DMSP/OLS夜间灯光数据,VIIRS传感器采用星上定标,可以得到更高精度的数据[15],新型夜间灯光数据空间分辨率的提高(500 m)也为小尺度级别的数据空间化研究提供了技术支持。目前有关NPP/VIIIRS夜间灯光的研究主要集中在模拟GDP[8,16-17]、电力消耗估算[18]、油气平台识别[19]等经济领域,而在人口空间化方面应用的较少。高义等[20]对比分析了2种夜间灯光数据在我国沿海地区人口空间化结果,研究表明,NPP/VIIRS夜间灯光数据反演人口的精度要高于DMSP/OLS夜间灯光数据,但是得到的结果精度并不是很高。因此,如何有效地借助NPP/VIIRS夜间灯光数据得到更高精度的人口空间化结果还需进一步研究。
综上所述,本研究在对比了2种不同的夜间灯光数据(DMSP/OLS、NPP/VIIRS)与人口统计数据相关性的基础上,引入土地利用数据,以川渝地区为研究区,与县级常住人口统计数据进行逐步回归建模,并且使用乡镇级统计人口对人口空间化结果进行精度评价,期望建立更有效的人口空间化数据。

2 研究区概况与数据源

2.1 研究区概况

川渝地区地处青藏高原与长江中下游平原的过渡地带,位于东经97°22′~110°12′、北纬26°02′~34°18′之间,总面积为56.74万km2,平均海拔2500 m左右,地势西高东低,区域地貌主要有高原、盆地、山地、丘陵、平原等。受地貌和季风环流的影响,川渝地区气候主要表现为西部的高原大陆性气候和东部的亚热带湿润季风气候。2013年,川渝地区常住人口为1.11亿,辖区包括18个地级市、3个自治州,共计218个县级行政区。川渝地区的地理位置及行政区划如图1所示。
Fig. 1 Location and elevation distribution of Sichuan-Chongqing region

图1 川渝地区位置及其高程分布图

川渝地区是我国西部经济发展的增长极,同时也是丝绸之路经济带与长江经济带的交汇点,具有联动东西、带动南北的区位优势,在国家“一带一路”和长江经济带发展战略中扮演重要角色[21]。作为中国西部人口重要聚居地之一,川渝地区也是洪涝、冰雹、泥石流、山体滑坡等自然灾害的多发区[22]。再加上该地区复杂的地形地貌条件,造成该地区人地关系矛盾突出,因此研究川渝地区的人口分布,可以为该地区经济协调稳定发展以及灾害风险评估与救援提供技术支持。

2.2 数据源

(1)DMSP/OLS数据:本研究使用的DMSP/OLS夜间灯光数据来自于美国国家地球物理数据中心(NGDC)发布的2013年夜间非辐射定标平均稳态数据(stable_light.avg_vis data)(图2(a)),该数据经过了去云处理,并且消除了背景噪声及短时光数据(火山气体、森林火灾、极光等)。像元灰度值介于0-63之间。数据下载地址https://www.ngdc.noaa.gov/eog/dmsp/downloadV4composites.html。
Fig. 2 DMSP/OLS and NPP/VIIRS nighttime light data of Sichuan-Chongqing region in 2013

图2 2013年川渝地区DMSP/OLS夜间灯光和NPP/VIIRS夜间灯光修正数据

(2)NPP/VIIRS数据:美国新一代极轨运行卫星系统预备项目(National Polar-orbiting Operational Environmental Satellite System Preparatory Project, NPP)卫星携带的VIIRS传感器共有22个波段,其中白天/夜间波段(Day/Night Band, DNB)星下地面分辨率为375 m,灰度区间(16 bit)大于DMSP/OLS(6 bit),可以识别微弱的灯光源,对地表照明分布的刻画更为准确(图2(b))。但该数据没有经过处理,还存在森林火灾、极光、火山等短时光数据和山顶积雪、干涸床等背景噪声。数据下载地址https://www.ngdc.noaa.gov/eog/viirs/download_monthly.html。
(3)土地利用与土地覆被数据(LUC数据):本次研究使用的LUC数据来自于中国科学院地理科学与资源研究所刘纪远团队[23],时间为2013年,该数据集将土地利用/覆被分为耕地、林地、草地、水域、建设用地和未利用6个一级类和有林地、灌木林、疏林地、其他林地以及高、中、低覆盖草地等22个二级类型(图3(a))。
Fig. 3 LUC and population density at county level of Sichuan-Chongqing region

图 3 川渝地区土地利用类型图与县级人口密度分布图

(4)人口统计数据:本次研究使用的人口数据是常住人口统计数据,主要来自《2014年四川省统计年鉴》、《2014年重庆市统计年鉴》(图3(b))。常住人口是指本行政区域内以下4部分人口:居住在在本行政区域,户口在本行政区域或户口在外县(市),离开户口登记地半年以上的人口;户口在本行政区域,居住在外县(市),离开户口登记地不到半年的人口;户口在本行政区域,居住在港澳台或国外的人口。
(5)行政区划边界数据:本次研究所使用的边界包括省、市、县、乡镇四级行政区划,数据来自于国家基础地理信息中心2012年发布的1:25万的行政区划边界(表1)。
Tab. 1 Data sources

表 1 数据来源表

数据类型 年份 数据来源 比例尺/分辨率
行政区划边界 2012 国家基础地理信息中心 1:25万
人口统计资料 2013 统计年鉴以各地区统计公报 县、乡镇
LUC 2013 中科院资源环境科学数据中心 1:10万/1 km
DMSP/OLS 2013 美国国家地球物理数据中心 1 km(采样后)
NPP/VIIRS 2013 美国国家地球物理数据中心 0.5 km(采样后)

2.3 数据预处理

首先对原始夜间灯光影像进行剪切、投影、重采样处理,采用双线性插值法分别将DMSP/OLS和NPP/VIIRS夜间灯光数据的空间分辨率重采样为 1 km和500 m,2种夜间灯光数据均转换为Albers等积圆锥投影(大地基准为WGS-84,中央经线为105°E,双标准纬线分别为25°N和47°N,起始原点为0°)。
NPP/VIIRS年数据合成:为消除偶然误差采用平均值法求得到年平均夜间灯光数据。具体计算公式如下:
A_R = 1 n i = 1 12 R i (1)
式中:A_R为夜间灯光反射率平均值;Rii月夜间灯光反射率。由于6月的夜间灯光数据缺失非常严重,因此不参与计算,其中: i≠6,n=11。
NPP/VIIRS背景噪声去除:合成后的年平均数据存在大量的背景噪声,本研究结合前人已有的研究[17,18],首先提取2013年的DMSP/OLS夜间灯光数据像元值非零区域作为掩膜,将掩膜区域以外的区域作为噪声部分,提取NPP/VIIRS夜间灯光数据;然后针对此数据,选择重庆市和成都市夜间灯光强度最高值作为有效灯光强度阈值,采用八领域算法[18]对夜间灯光数据进行平滑处理,最终得到川渝地区NPP/VIIRS夜间灯光修正数据。

3 分析方法

3.1 人口空间化方法

在SPSS软件下,以各区县常住人口统计数据作为因变量,各区县内不同土地利用类型下的夜间灯光亮元数、暗元数及灯光总亮度为自变量(影像中亮度值为0的像元作为暗元,亮度不为0的像元作为亮元),进行逐步回归分析,得到模型自变量及回归系数,最后根据建立的模型对人口进行空间化。考虑到分别提取掩膜进行计算过程较为复杂,为方便计算,本研究采用格网计算的方法,对LUC数据和夜间灯光数据进行叠加,具体方法为:首先分别建立川渝地区500 m和1000 m的格网矢量数据,以此来统计各格网上的对应的土地利用类型、亮元数、暗元数、灯光总亮度;然后与县级行政区划边界进行叠加分析,得到各县各土地利用类型上的灯光亮元数、暗元数、灯光总亮度。人口空间化流程如图4所示。
Fig. 4 Spatial distribution of population based on nighttime light data and LUC

图 4 基于夜间灯光数据和土地利用数据的人口空间化流程图

3.2 逐步回归建模

提取的模型自变量包括:NU是暗元数,代表一种土地利用类型下无灯光区面积;NL是亮元数,为同种土地利用类型下有灯光区面积;LE代表该种土地利用类型下灯光总辐射亮度值。模型的构建采用逐步回归法,各自变量进入方程的置信水平为0.05,剔除方程的置信水平为0.1。回归方程的常量为正。初次建立的模型可能会出现某些自变量系数为负的情况,主要原因是各变量之间存在共线性的问题,使模型估算出来的人口可能会出现负值,与实际情况不符合。在本研究中,直接剔除系数为负的变量,然后将剩余变量再次引入模型,最终进入模型的自变量的系数全为正,且常量为正。模型表达式为:
P i = P 0 + j = 1 M a j × N U ij + b j × N L ij + c j × L E ij (2)
式中:Pi为第i个县级统计人口;P0为常数;M为土地类型数;NUijNLijLEij分别为第i个县第j种土地利用类型上的亮元数、暗元数和总亮度指数;aj、bj、cj为回归系数,利用上述结果得到像元尺度人口:
P ijk = P 0 N i + j = 1 M a j × N U ijk + b j × N L ijk + c j × L E ijk (3)
式中:Pijk为第i个县内第j种土地利用类型上第k个像元上的人口数;Ni为第i个县内像元个数;M为土地利用类型数;ajbjcj为回归系数;NUijkNLijkLEjjk分别为第i个县第j种土地利用类型第k个像元上的亮元数、暗元数和灯光总亮度。
P ijk ' = P ijk × P i ¯ P i (4)
式中: P ijk ' 为最终栅格人口; P i ¯ 为第i县统计人口; P i 为第i县所有像元值之和。
式(2)是县域单元上人口-各类型土地灯光强度关系的表达式,通过该公式,将可以反推得到相关变量(各类型土地暗元数、亮元数、灯光强度)的回归系数;式(3)是将式(2)在县域单元上得到的各变量回归系数纳入到计算公式中,开展像元尺度上的人口估算;式(4)是利用县域单元上的人口普查统计数据与基于式(3)得到的县域单元上全部像元上的人口空间统计数进行对比,从而对各像元上的人口进行微调,从而确保县域尺度上汇总的人口空间分布模拟数据与实际的统计数据保持一致。

3.3 精度评价

对于模拟得到的人口需要进行精度评价和误差分析,本研究选取了相关系数(R)、均方根误差(Root Mean Square Error, RMSE)、平均绝对误差(Mean Absolute Error,MAE)、平均相对误差(Mean relative Error,MRE)来进行评价。具体公式如下:
R = i = 1 n P i - P ̅ P E i - PE ¯ i = 1 n P i - P ̅ 2 i = 1 n P E i - PE ¯ 2 (5)
RMSE = i = 1 n P E i - P i 2 n (6)
MAE = 1 n i = 1 n P E i - P i (7)
MRE = 1 n i = 1 n P E i - P i P i (8)
R E i = P E i - P i P i (9)
式中:Pi代表i行政单元内统计人口数;PEi表示i行政单元内人口的估计数;n代表行政单元的个数; P ̅ 代表统计人口数平均值; PE ¯ 表示人口的估计数平均值。

4 结果与讨论

4.1 人口与夜间灯光的相关性

图5可以看出,川渝地区3种模型拟合人口密度和DMSP/OLS夜间灯光数据按绝对系数大小的模型顺序为:幂函数模型、多项式模型、线性模型,绝对系数(R2)分别为0.85、0.77、0.58,相关系数(R)分别为:0.924、0.878、0.765;对于NPP/VIIRS夜间灯光数据来说,3种模型拟合结果与DMSP/OLS夜间灯光数据一致,绝对系数(R2)分别为0.86、0.75、0.75,相关系数(R)分别为0.926、0.866、0.864。通过对比可以看出,NPP夜间灯光数据拟合人口的能力要优于DMSP/OLS夜间灯光数据,而且拟合结果受模型的影响不大。
Fig. 5 Correlation between population density at county level and mean nighttime light data

图5 川渝地区区县常住人口密度与平均夜间灯光指数相关关系 ||||注:黑线是线性回归模型;红线是多项式回归模型;蓝线是幂函数模型

4.2 人口与LUC的相关性

图6可以看出,川渝地区主要土地利用类型 为林地(20.1万km2)、草地(17.2万km2)和耕地 (15.6万km2)3类,其分布与地形密切相关。由图3可以看出,耕地主要分布在中东部的盆地地区,草地主要分布在西北高海拔地区,林地主要分布在西南和东部中低海拔地区。城乡、工矿、建设用地面积为0.71万km2,主要集中在成都市和重庆市及其周围区县。进一步,通过统计川渝地区各土地利用类型面积,并在SPSS中分别计算各区县常住人口与各土地利用类型覆被面积的相关性,分析结果表明土地利用分布与人类活动关系密切。由表2可以看出,各类型土地利用数据与人口的相关性强弱依次为耕地、城乡工矿居民用地、草地、林地、未利用土地、水域,相关系数分别为0.62、0.57、-0.45、-0.44、-0.3、0.08。其中,耕地和城乡工矿居民用地与人口呈现显著(P=0.01)正相关关系,林地、草地、未利用土地与人口呈现显著(P=0.01)负相关关系。在使用土地利用数据参与人口空间化研究过程中,考虑人口分布的实际情况,水域和未利用土地类型不参与人口空间化计算。
Fig. 6 Land use area of Sichuan-Chongqing region in 2013(104km2

图 6 2013年川渝地区土地利用面积图(104km2

Tab. 2 Correlation analysis between land use and population in Sichuan-Chongqing region

表 2 川渝地区各土地利用与人口的相关性分析

耕地 林地 草地 水域 城乡、工矿、居民用地 未利用土地
相关性(R) 0.62** -0.44** -0.45** 0.08 0.57** -0.30**

注:**表示在0.01置信水平上显著

4.3 人口空间化结果

4.3.1 模型参数
表3可以看出,2种夜间灯光数据和土地利用叠加后最终进入模型的变量相同,包括耕地亮元数、耕地暗元数、林地灯光总亮度、城镇及建设用地灯光总亮度,且都通过了显著性检验(p=0.01)。在实际情况中,人口只分布于城镇及建设用地上,本文建模过程中,考虑了耕地和林地是由于基于卫星遥感解译的土地利用产品精度问题、图斑上图标准问题,在耕地、林地、草地、甚至荒漠、水域中,都有可能存在零星分布农村居民点、农牧民独立房屋、船舍、帐篷、蒙古包、毡房等设施。上述微小、零星分布、但数量众多的居住设施在本研究所使用的1:10万LUC数据中是无法体现,但又确实存在的,因此本研究对上述土地类型中的人口分布可能性赋予了一定的权重。据了解,中国科学院地理科学与资源研究所其他相关课题组、国家发改委宏观院等相关机构在制作同类地图时,也有同类型的考虑[24]。基于DMSP/OLS夜间灯光数据和LUC建立的模型的复相关系数为0.796,而基于NPP/VIIRS夜间灯光数据和LUC建立的模型的复相关系数为0.817。
Tab. 3 Regression coefficients of the model

表3 模型回归系数表

土地利用类型 DMSP/OLS NPP/VIIRS
系数 Sig. 系数 Sig.
耕地 NU 392.399 0.000 95.736 0.000
NL 442.249 0.000 163.979 0.000
林地 LE 29.270 0.024 85.934 0.028
城镇及建设用地 LE 106.833 0.000 71.042 0.000
常数 Con 39 848.898 0.050 45 683.36 0.017
R2 0.796 0.817
4.3.2 川渝地区人口空间化结果
图7(a)、(b)显示了基于2种夜间灯光数据和LUC数据的川渝地区人口空间化结果。为方便直观地对比2种空间化结果,本文对500 m人口空间化结果的单位进行了换算,将基于NPP/VIIRS 500 m分辨率影像所得人口空间密度估算成果的单位转化为人/km2,可以看出,得到的人口空间分布情况大致相同,人口主要集中在居民地和城镇建设用地上,各区县的人口密度高值区主要集中在县城所在地,其中重庆市和成都市人口最集中,密度最高。川渝地区的常住人口分布是以重庆市和成都市为2个中心分布,重庆市是自治区,经济发展较快,成都市是四川省的省会城市,城市化进程明显,人口大量聚集,人口密度极高。
Fig. 7 Spatial distribution of population in Sichuan-Chongqing region in 2013

图 7 2013年川渝地区人口空间化结果

对比重庆市人口空间化结果(图7(c)、(e))可以发现,2种模拟结果大致空间分布格局相同,即人口密度呈现中心城区高,四周低的分布情况。人口密度最高的区域为渝中区,1 km2人口超过25 000人,与实际情况(28 515.2人/km2)相符。对于成都市(图7(d)、(f))来说,可以看出高密度人口区域主要集中在金牛、青羊、武侯、成华、锦江5个区,周围人口密度较低,同样的基于NPP/VIIRS夜间灯光数据得到的人口空间化结果分辨率比基于DMSP/OLS夜间灯光数据得到的结果高,能反映出同一区县内人口分布的变化情况。同时,中心城区与周边城区人口密度变化更加自然,更符合实际人口分布情况。

4.4 精度检验

随机选取500个乡镇级别的2013年常住人口统计数据作为真实人口数据,利用这500个乡镇区划统计模拟的2种人口空间化结果作为估计值;分别计算整体的平均绝对误差(MAE)、平均相对误差(MRE)、均方根误差(RMSE),另外计算各个乡镇的人口估计相对误差(RE),并分级统计分析。
表4表示的是基于2种夜间灯光数据和LUC数据模拟得到的空间分布人口的3种误差指标:MAE、MRE、RMSE。由表4可以看出,基于NPP/VIIRS夜间灯光数据得到的人口空间误差在3种误差上都要小于DMSP/OLS夜间灯光数据,具体表现为:① 在MAE方面,基于DMSP/OLS夜间灯光数据得到的误差为10 851,而基于NPP/VIIRS夜间灯光数据得到的误差为10 450;② 在MRE方面,基于DMSP/OLS夜间灯光数据得到的误差为46.3%,而基于NPP/VIIRS夜间灯光数据得到的误差为44.62%;③ 在RMSE方面,基于DMSP/OLS夜间灯光数据得到的误差为637 518,而基于NPP/VIIRS夜间灯光数据得到的误差为587 170。对比发现,基于NPP/VIIRS夜间灯光数据得到的人口空间化结果精度要高于DMSP/OLS夜间灯光数据。
Tab. 4 Statistics of spatial distribution of population errors

表 4 人口空间化误差统计表

误差类型 DMSP/OLS NPP/VIIRS
MAE 10 851 10 450
MRE/% 46.3 44.62
RMSE 637 518 587 170
表5是对500个乡镇级别的人口的相对误差做分级统计,图8表示的是各级别相对误差占比。可以看出基于NPP/VIIRS夜间灯光数据得到的人口空间化精度要高于DMSP/OLS。具体表现为:基于NPP/VIIRS夜间灯光数据得到的人口空间化结果出现严重高估的乡镇个数为85,占总乡镇数的17%;基于DMSP/OLS夜间灯光数据得到的人口空间化结果出现严重高估的乡镇个数为76个,占比为15%,相比前者少2%;基于NPP/VIIRS夜间灯光数据得到的人口空间化结果一般高估的乡镇个数为83个,占比17%,基于DMSP/OLS夜间灯光数据得到的人口空间化结果一般高估的乡镇个数为94个,相比NPP/VIIRS多11个,占比为19%;在准确估计的乡镇中,基于NPP/VIIRS夜间灯光数据得到的人口空间化结果为188,相比DMSP/OLS多22个(166),在一般低估的乡镇中,基于NPP/VIIRS夜间灯光数据得到的人口空间化误差个数为116,比DMSP/OLS少13个,在严重低估的乡镇中,基于NPP/VIIRS夜间灯光数据得到的人口结果比DMSP/OLS少7个。总体来看,基于NPP/VIIRS夜间灯光数据得到的人口空间化结果精度要高于DMSP/OLS。
Tab. 5 Statistics of relative error classification

表 5 相对误差分级统计表

夜间灯光 严重低估 一般低估 较准确估计 一般高估 严重高估
(-100%,-50%] (-50%,-20%] (-20%,20%] (20%,50%] (50%,100%]
DMSP/OLS 35 129 166 94 76
NPP/VIIRS 28 116 188 83 85
Fig. 8 Statistics of relative error ratios

图 8 相对误差占比统计图

通过与已有的研究对比发现,高义等[20]基于2种夜间灯光数据得到的空间化人口中能够较准确地估计的乡镇个数占比为23.75%(DMSP/OLS)和26.75%(NPP/VIIRS),远低于本研究中的33%和38%。这表明在使用夜间灯光数据进行人口空间化时引入土地利用数据可以大大提高人口空间化的精度。此外,分析出现明显高估和明显低估的原因可能与该地区的气候、海拔等其他影响人类分布的因素有关,而且这些影响因素与人口大多是非线性的关系,因此今后可以从增加与人口分布相关影响因子和改变建模方法这2个方面入手以进一步的提高人口空间化的精度。

5 结论

人口空间分布直观反映了人类活动的范围和强度,是研究和表征人地关系的有效指标和必需数据。夜间灯光数据是人口空间化过程中重要的数据源。本研究对比分析了2种夜间灯光数据在模拟人口时的适宜性,同时结合土地利用数据对川渝地区常住人口统计数据进行了空间化,并对人口空间化结果进行了精度验证。本文的主要结论如下:
(1)在拟合人口时,NPP/VIIRS的效果整体要好于DMSP/OLS。具体表现在使用不同的模型拟合时,人口与NPP/VIIRS的绝对系数均高于0.75,说明NPP/VIIRS拟合人口时受模型的波动较小。
(2)各类型土地利用数据与人口的相关性强弱依次为耕地、城乡工矿居民用地、草地、林地、未利用土地、水域,相关系数分别为0.62、0.57、-0.45、-0.44、 -0.3、0.08。建模结果表明除了建设用地,其他类型用地(耕地、林地)对人口分布会有影响,在对人口进行建模时不应该只考虑建成区。
(3)在2种夜间灯光数据分别与LUC结合时,使用DMSP/OLS夜间灯光数据和NPP/VIIRS夜间灯光数据得到的逐步回归模型的复相关系数R2分别为0.796、0.817,且进入模型的变量的回归系数都经过了正数检验和显著性检验。模型拟合率较高,基于DMSP/OLS得到的人口空间化结果分辨率为1 km,基于NPP/VIIRS夜间灯光数据得到的人口空间化结果分辨率为500 m,相比DMSP/OLS夜间灯光数据提高了一倍。
(4)在夜间灯光数据与LUC数据结合得到的川渝地区人口空间化结果中,使用NPP/VIIRS夜间灯光数据得到的结果精度要高于DMSP/OLS夜间灯光数据,具体表现为:① 在MAE方面,基于NPP/VIIRS夜间灯光数据得到的结果误差为10 450,基于DMSP/OLS夜间灯光数据得到的结果误差为10 851;② 在MRE方面,基于NPP/VIIRS夜间灯光数据得到的结果误差为44.62%,基于DMSP/OLS夜间灯光数据得到的结果误差为46.3%;③ 在RMSE方面,基于NPP/VIIRS夜间灯光数据得到的结果误差为587 170,基于DMSP/OLS夜间灯光数据得到的结果误差为637 518;④ 在相对误差方面,基于NPP/VIIRS夜间灯光数据得到的结果较准确估计的乡镇个数(188个)要多于DMSP/OLS(166个),在相对误差较大的等级中,除严重高估的乡镇个数(85个)略高于DMSP/OLS(76个),其余误差较大的乡镇中,NPP/VIIRS夜间灯光数据都要少于DMSP/OLS。

The authors have declared that no competing interests exist.

[1]
林丽洁,林广发,颜小霞,等.人口统计数据空间化模型综述[J].亚热带资源与环境学报,2010,5(4):10-16.人口统计数据空间化是人口信息与其他资源环境、社会经济等信息进行空间集成的基础.本研究对国内外人口统计数据空间化研究进行总结,归纳了水热条件、地形地貌、土地利用、交通廊道、夜间灯光等不同建模参考因素对人口空间分布的影响,并分析比较了10个主要的人口统计数据空间化模型,进而对当前人口统计数据空间化研究中存在的问题做了总结,并讨论未来的研究方向.综述认为人口统计数据空间化的研究将向数据获取多源化、建模因素综合化、模拟格网精细化、模型应用实用化等方向发展;目前需要改进的问题包括:1)统一的人口数据统计标准;2)人口结构特性相关的空间化,特别是流动人口的空间分布特征识别;3)城市街区尺度的空间化方法研究;4)多源数据与人口动态信息综合中的时相匹配;5)统一的空间化指标量化方法;6)模型参数优化与精度验证方法完善.

DOI

[Lin L J, Lin G F, Yan X X, et al.A review on spatialization models of census data[J]. Journal of Subtropical Resources and Environmen, 2010,5(4):10-16. ]

[2]
Alahmadi M, Atkinson P, Martin D.Estimating the spatial distribution of the population of Riyadh, Saudi Arabia using remotely sensed built land cover and height data[J]. Computers, Environment and Urban Systems, 2013,41:167-176.This paper investigates the use of Landsat ETM+, remotely sensed height data, ward-level census population, and dwelling units to downscale population in Riyadh, Saudi Arabia. Regression analysis is used to model the relationship between density of dwelling units and built area proportion at the block level and the coefficients used to downscale density of dwelling units to the parcel level. The population distribution is estimated based on average population per dwelling unit. Seven models were fitted and compared. First, a conventional approach, using ISODATA-classified built land cover alone as a covariate, is used as a benchmark against which to evaluate six more sophisticated models. The conventional model results in low accuracy measured by overall relative error (ORE) (+116%). Approaches for potentially increasing accuracy are explored, incorporating above-surface height data into the downscaling process. These include masking out zero and near-zero height areas when estimating built area; using height to estimate the number of floors; replacing the ISODATA model with a support vector machine; estimating volume-adjusted habitable space; stratifying the study area into different building categories; and preservation of the dependent variable for the best model. These approaches result in large increases in accuracy in the density of dwelling unit estimates. However, while the height data accounts for the vertical dimension (primarily through the number of floors), it is not possible to account for variation in dwelling density which arises due to other factors such as living standards, affluence and other spatially varying factors, without further data.

DOI

[3]
叶宇,刘高焕,冯险峰.人口数据空间化表达与应用[J].地球信息科学,2006,8(2):59-65.

[Ye Y, Liu G H, Feng X F.Presentation and application of spatial distribution of the population data[J]. Geo-information Science, 2006,8(2):59-65. ]

[4]
胡云锋,王倩倩,刘越,等.国家尺度社会经济数据格网化原理和方法[J].地球信息科学学报,2011,13(5):573-578.空间信息获取和处理技术,可将传统关系型社会经济数据转变为空间化的矢量或者格网数据。首先,分析了传统关系型社会经济数据库的不足,并对社会经济数据格网化理论研究及格网数据库建设的历史做了回顾;在明确社会经济数据格网化基本概念的基础上,提出了社会经济数据格网化的3个基本要求,即时间可比、空间一致和逻辑自洽;同时提出了一个包括24个关键指标的国家尺度社会经济格网数据库的指标体系,认为社会经济格网数据库生产过程的主要步骤为逻辑检查、空间匹配、代码匹配、空间离散和检查校验;研究对国家尺度社会经济指标的空间离散过程和离散模型、不同层级社会经济数据的整合和离散策略进行了重点分析。研究最后就社会经济数据格网化过程中存在的主要问题进行了总结。

DOI

[Hu Y F, Wang Q Q, Liu Y, et al.Principles and methods of the social and economic grid database at national scale[J]. Journal of Geo-information Science,2011,13(5):573-578. ]

[5]
柏中强,王卷乐,杨飞.人口数据空间化研究综述[J].地理科学进展,2013,32(11):1692-1702.人口数据空间化研究旨在发掘和展现人口统计数据中隐含的空间信息,并以地理格网或其他区域划分的形式再现客观世界的人口分布,具有重要的科学意义。人口空间分布数据有助于从不同地理尺度和地理维度对人口统计数据形成有益补充,其应用广泛,相关研究方兴未艾。主要从以下3 个方面对人口数据空间化研究进行综述:① 主要空间化方法的原理及其适用性;② 空间化中用到的建模参考因素,并结合具体应用案例分析其作用机理;③ 典型人口空间化数据集。在此基础上,分析了现阶段人口数据空间化所运用的输入数据的质量和详细程度、尺度效应及时空分辨率、长时间序列数据集和精度检验等方面存在的问题;并探讨了人口数据空间化未来的研究方向。

DOI

[Bai Z Q, Wang J L, Yang F.Research progress in spatialization of population data[J]. Progress in Geography, 2013,32(11):1692-1702. ]

[6]
Azar D, Engstrom R, Graesser J, et al.Generation of fine-scale population layers using multi-resolution satellite imagery and geospatial data[J]. Remote Sensing of Environment, 2013,130:219-232.78 Image classification and dasymetric mapping produce a detailed population layer. 78 Technique distributes third level population counts to 100-meter pixels. 78 Areal estimates of population compared to LandScan, GPW, and GRUMP.

DOI

[7]
Lung T, Lübker T, Ngochoch J K, et al.Human population distribution modelling at regional level using very high resolution satellite imagery[J]. Applied Geography, 2013,41:36-45.Modelling the distribution of human population based on satellite-derived information has become an important field of research, providing valuable input e.g. for human impact assessments related to the management of threatened ecosystems. However, few regional-scale studies have been conducted in developing countries, where detailed land cover data is usually absent, and the potential of very high resolution (VHR) satellite imagery in this context has not been explored yet. This study uses results obtained through object-based image analysis (OBIA) of QuickBird imagery for a subset of a highly populated rural area in western Kenya. Functions are established that approximate frequency distributions of QuickBird-derived locations of houses in relation to five factors. These factors are known to impact settlement patterns and data is available for the entire study area. Based on an overall probability coefficient (weight) calculated from the single functions, human population is redistributed at the smallest administrative level available (version A). In addition, the problem of artefacts remaining at administrative boundaries is addressed by combining the approach with the pycnophylactic smoothing algorithm ( Tobler, 1979 ) (version B). The results show distinct patterns of population distribution, with particular influence of rivers/streams and slope, while version B in addition is free of boundary artefacts. Despite some limitations compared to models based on detailed land cover data (e.g. the ability of capturing abrupt changes in population density), a visual and numerical evaluation of the results shows that using houses as classified from VHR imagery for a study area subset works well for redistributing human population at the regional level. This approach might be suitable to be applied also in other regions of e.g. sub-Saharan Africa.

DOI

[8]
柴子为,王帅磊,乔纪纲.基于夜间灯光数据的珠三角地区镇级GDP估算[J].热带地理,2015,35(3):379-385.<p>高精度地表GDP分布数据是开展小尺度区域发展相关研究的关键数据,但通常难以获得完整数据序列。文章比较了DMSP-OLS和NPP-VIRS两种夜间灯光数据在小尺度单元GDP估算工作上的适应性,证实NPP-VIIRS数据在镇级GDP估算中有更好的性能。利用修正后的NPP-VIIRS夜间灯光数据建立珠三角地区镇级GDP估算模型,并通过地区差异系数对估计结果进行校正。对2013年镇级GDP估算的实验结果总体精度达到85%。估算结果能够用于填补珠三角地区部分镇级GDP统计数据缺失,为相关研究获取、对比GDP数据提供技术手段。</p>

[Chai Z W, Wang S L, Qiao J G.Township GDP estimation of the Pearl River Delta based on the NPP-VIIRS night-time satellite data[J]. Tropical Geography, 2015,35(3):379-385. ]

[9]
韩向娣,周艺,王世新,等.夜间灯光遥感数据的GDP空间化处理方法[J].地球信息科学学报,2012,14(1):128-136.随着夜间灯光遥感数据的应用日渐成熟和资源环境研究领域,对空间型社会经济数据的需求增加,利用相关分析和回归分析的方法,首次定量探讨夜间灯光数据与统计型的社会经济数据的空间关系。为提高模型精度,按照我国省级行政边界分区建模,分析全国县级的地区生产总值、第一产业、第二产业、第三产业分别与夜间灯光指数的空间相关关系,最终建立全国的1km GDP密度图。结果表明,全国范围的夜间灯光数据与第一产业的相关性不明显,相关系数0.554,模型拟合效果差,R<sup>2</sup>为0.306;夜间灯光数据与地区生产总值、第二产业、第三产业均有明显的对数线性关系,尤其是与第二产业和第三产业之和,相关系数为0.824,R<sup>2</sup>为0.679。利用分区模型估算,生成的GDP密度图能较完整地反映全国社会经济分布详况,以及宏观分布特征。

[Han X D, Zhou Y, Wang S X, et al.GDP spatialization in China based on nighttime imagery[J]. Journal of Geo-information Science, 2012,14(1):128-136. ]

[10]
杨洋,黄庆旭,章立玲.基于DMSP/OLS夜间灯光数据的土地城镇化水平时空测度研究——以环渤海地区为例[J].经济地理,2015,35(2):141-148,168.

[Yang Y, Huang Q X, Zhang L L.The spatial-temporal measurement on the land urbanization level using DMSP/OLS nighttime light data: A case study of Bohai rim[J]. Economic Geography, 2015,35(2):141-148,168. ]

[11]
Yang X, Yue W, Gao D.Spatial improvement of human population distribution based on multi-sensor remote-sensing data: An input for exposure assessment[J]. International Journal of Remote Sensing, 2013,34(15):5569-5583.A spatial mismatch of hazard data and exposure data (e.g. population) exists in risk analysis. This article provides an integrated approach for a rapid and accurate estimation of population distribution on a per-pixel basis, through the combined use of medium and coarse spatial resolution remote-sensing data, namely the Defense Meteorological Satellite Program Operational Linescan System (DMSP/OLS) night-time imagery, enhanced vegetation index (EVI), and digital elevation model (DEM) data. The DMSP/OLS night-time light data have been widely used for the estimation of population distribution because of their free availability, global coverage, and high temporal resolution. However, given its low-radiometric resolution as well as the overglow effects, population distribution cannot be estimated accurately. In the present study, the DMSP/OLS data were combined with EVI and DEM data to develop an elevation-adjusted human settlement index (EAHSI) image. The model for population density estimation, developed based on the significant linear correlation between population and EAHSI, was implemented in Zhejiang Province in southeast China, and a spatialized population density map was generated at a resolution of 250 m 250 m. Compared with the results from raw human settlement index (59.69%) and single night-time lights (35.89%), the mean relative error of estimated population by EAHSI has been greatly reduced (17.74%), mainly due to the incorporation of elevation information. The accurate estimation of population density can be used as an input for exposure assessment in risk analysis on a regional scale and on a per-pixel basis.

DOI

[12]
Small C, Pozzi F, Elvidge C D.Spatial analysis of global urban extent from DMSP-OLS night lights[J]. Remote Sensing of Environment, 2005,96(3-4):277-291.Previous studies of DMSP-OLS stable night lights have shown encouraging agreement between temporally stable lighted areas and various definitions of urban extent. However, these studies have also highlighted an inconsistent relationship between the actual lighted area and the boundaries of the urban areas considered. Applying detection frequency thresholds can reduce the spatial overextent of lighted area (“blooming”) but thresholding also attenuates large numbers of smaller lights and significantly reduces the information content of the night lights datasets. Spatial analysis of the widely used 1994/1995 stable lights data and the newly released 1992/1993 and 2000 stable lights datasets quantifies the tradeoff between blooming and attenuation of smaller lights. For the 1992/1993 and 2000 datasets, a 14% detection threshold significantly reduces blooming around large settlements without attenuating many individual small settlements. The corresponding threshold for the 1994/1995 dataset is 10%. The size–frequency distributions of each dataset retain consistent shapes for increasing thresholds while the size–area distributions suggest a quasi-uniform distribution of lighted area with individual settlement size between 10 and 1000 km equivalent diameter. Conurbations larger than 80 km diameter account for 0290% can often reconcile lighted area with built area in the 1994/1995 dataset but there is not one threshold that works for a majority of the 17 cities considered. Even 100% thresholds significantly overestimate built area for the 1992/1993 and 2000 datasets. Comparison of lighted area with blooming extent for 10 lighted islands suggests a linear proportionality of 1.25 of lighted to built diameter and an additive bias of 2.7 km. While more extensive analyses are needed, a linear relationship would be consistent with a physical model for atmospheric scattering combined with a random geolocation error. A Gaussian detection probability model is consistent with an observed sigmoid decrease of detection frequency for settlements <0210 km diameter. Taken together, these observations could provide the basis for a scale-dependent blooming correction procedure that simultaneously reduces geolocation error and scattering induced blooming.

DOI

[13]
Sutton P, Roberts D, Elvidge C, et al.A comparison of nighttime satellite imagery and population density for the Continental United Sates[J]. American Society for Photogrammetry and Remote Sensing, 1997,63(11):1303-1313.ABSTRACT The striking apparent correlation between nighttime satellite imagery and human population density was explored for the continental United States. The nighttime stable-lights im-agery was derived from the visible near-ZR band of 231 orbits of the Defense Meteorological Satellite Program Operational Linescan System (DMSP-OLS). The population density data were generated from a gridded vector dataset of the 1992 United States census block group polygons. Both datasets are at a one-square-kilometre resolution. The two images were co-registered and correlation between them was measured at a range of spatial scales, including aggregation to state and county levels. DMSP imagery showed strong correlations at aggregate scales, and analysis of the saturated areas of the images showed strong correlations between the areas of satu-rated clusters and the populations those areas cover. The non-zero pixels of the DMSP imagery correspond to only 10 percent of the land cover yet account for over 80 percent of the continental United States population. Spatial analysis of the clusters of the saturated pixels predicts population with an R2 of 0.63. Consequently, the DMSP imagery may prove to be useful to inform a "smart interpolation" program to im-prove maps and datasets of human population distributions in areas of the world where good census data may not be available or do not exist.

DOI

[14]
Zhang Q, Seto K C.Mapping urbanization dynamics at regional and global scales using multi-temporal DMSP/OLS nighttime light data[J]. Remote Sensing of Environment, 2011,115(9):2320-2329.78 We map urbanization dynamics at regional and global scales with nighttime light data. 78 Differences in urbanization trajectories can be identified using temporal signatures. 78 We use an iterative clustering method to distinguish stable urban from urban growth. 78 From 1992 through 2000 India experienced higher rates of urbanization than China. 78 From 2000 through 2008 China experienced higher rates of urbanization than India.

DOI

[15]
Baugh K, Hsu F-C, Elvidge C D, et al.Nighttime lights compositing using the VIIRS day-night band: Preliminary results[J]. Proceedings of the Asia-Pacific Advanced Network, 2013,35:70-86.Dramatically improved nighttime lights capabilities are presented by the launch of the National Polar-orbiting Partnership (NPP) Visible Infrared Imaging Radiometer Suite (VIIRS) Day Night Band (DNB) sensor. Building on 18 years of experience compositing nighttime data from the Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS), NOAA NGDC Earth Observation Group has started adapting their algorithms to process these new data. The concept of compositing nighttime data comprises combining only high quality data components over a period of time to improve sensitivity and coverage. For this work, flag image are compiled to describe image quality. The flag categories include: daytime, twilight, stray light, lunar illuminance, noisy edge of scan data, clouds, and no data. High quality data is defined as not having any of these attributes present. Two methods of reprojection are necessary due to data collection characteristics. Custom algorithms have been created to terrain-correct and reproject all data to a common 15 arc second grid. Results of compositing over two time periods in 2012 are presented to demonstrate data quality and initial capabilities. These data can be downloaded at http://www.ngdc.noaa.gov/eog/viirs/download_viirs_ntl.html .

DOI

[16]
Dai Z, Hu Y, Zhao G.The suitability of different nighttime light data for GDP estimation at different spatial scales and regional levels[J]. Sustainability, 2017,9(2):305-320.Nighttime light data offer a unique view of the Earth surface and can be used to estimate the spatial distribution of gross domestic product (GDP). Historically, using a simple regression function, the Defense Meteorological Satellite Program Operational Linescan System (DMSP/OLS) has been used to correlate regional and global GDP values. In early 2013, the first global Suomi National Polar-orbiting Partnership (NPP) visible infrared imaging radiometer suite (VIIRS) nighttime light data were released. Compared with DMSP/OLS, they have a higher spatial resolution and a wider radiometric detection range. This paper aims to study the suitability of the two nighttime light data sources for estimating the GDP relationship between the provincial and city levels in Mainland China, as well as of different regression functions. First, NPP/VIIRS nighttime light data for 2014 are corrected with DMSP/OLS data for 2013 to reduce the background noise in the original data. Subsequently, three regression functions are used to estimate the relationship between nighttime light data and GDP statistical data at the provincial and city levels in Mainland China. Then, through the comparison of the relative residual error (RE) and the relative root mean square error (RRMSE) parameters, a systematical assessment of the suitability of the GDP estimation is provided. The results show that the NPP/VIIRS nighttime light data are better than the DMSP/OLS data for GDP estimation, whether at the provincial or city level, and that the power function and polynomial models are better for GDP estimation than the linear regression model. This study reveals that the accuracy of GDP estimation based on nighttime light data is affected by the resolution of the data and the spatial scale of the study area, as well as by the land cover types and industrial structures of the study area.

DOI

[17]
Li X, Xu H, Chen X, et al.Potential of NPP-VIIRS nighttime light imagery for modeling the regional economy of China[J]. Remote Sensing, 2013,5(6):3057-3081.Historically, the Defense Meteorological Satellite Program Operational Linescan System (DMSP-OLS) was the unique satellite sensor used to collect the nighttime light, which is an efficient means to map the global economic activities. Since it was launched in October 2011, the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor on the Suomi National Polar-orbiting Partnership (NPP) Satellite has become a new satellite used to monitor nighttime light. This study performed the first evaluation on the NPP-VIIRS nighttime light imagery in modeling economy, analyzing 31 provincial regions and 393 county regions in China. For each region, the total nighttime light (TNL) and gross regional product (GRP) around the year of 2010 were derived, and a linear regression model was applied on the data. Through the regression, the TNL from NPP-VIIRS were found to exhibit R2 values of 0.8699 and 0.8544 with the provincial GRP and county GRP, respectively, which are significantly stronger than the relationship between the TNL from DMSP-OLS (F16 and F18 satellites) and GRP. Using the regression models, the GRP was predicted from the TNL for each region, and we found that the NPP-VIIRS data is more predictable for the GRP than those of the DMSP-OLS data. This study demonstrates that the recently released NPP-VIIRS nighttime light imagery has a stronger capacity in modeling regional economy than those of the DMSP-OLS data. These findings provide a foundation to model the global and regional economy with the recently availability of the NPP-VIIRS data, especially in the regions where economic census data is difficult to access.

DOI

[18]
Shi K, Yu B, Huang Y, et al.Evaluating the ability of NPP-VIIRS nighttime light data to estimate the gross domestic product and the electric power consumption of China at multiple scales: A comparison with DMSP-OLS data[J]. Remote Sensing, 2014,6(2):1705-1724.The nighttime light data records artificial light on the Earth’s surface and can be used to estimate the spatial distribution of the gross domestic product (GDP) and the electric power consumption (EPC). In early 2013, the first global NPP-VIIRS nighttime light data were released by the Earth Observation Group of National Oceanic and Atmospheric Administration’s National Geophysical Data Center (NOAA/NGDC). As new-generation data, NPP-VIIRS data have a higher spatial resolution and a wider radiometric detection range than the traditional DMSP-OLS nighttime light data. This study aims to investigate the potential of NPP-VIIRS data in modeling GDP and EPC at multiple scales through a case study of China. A series of preprocessing procedures are proposed to reduce the background noise of original data and to generate corrected NPP-VIIRS nighttime light images. Subsequently, linear regression is used to fit the correlation between the total nighttime light (TNL) (which is extracted from corrected NPP-VIIRS data and DMSP-OLS data) and the GDP and EPC (which is from the country’s statistical data) at provincial- and prefectural-level divisions of mainland China. The result of the linear regression shows that R2 values of TNL from NPP-VIIRS with GDP and EPC at multiple scales are all higher than those from DMSP-OLS data. This study reveals that the NPP-VIIRS data can be a powerful tool for modeling socioeconomic indicators; such as GDP and EPC.

DOI

[19]
李强,苏奋振,王雯玥.基于VIIRS数据的油气平台提取技术研究[J].地球信息科学学报,2017,19(3):398-406.油气平台作为海上油气资源勘探开发的主要设备之一,其数量与空间分布反映了一个区域油气资源的开发状况。普通光学影像易受天气状况影响,而雷达数据成本较高,这给油气平台的检测带来了一定的困难。由于油气平台作业过程中,需要灯光照明,同时伴气的燃烧也产生很强的灯光,因此可以通过检测灯光来实现油气平台的提取。针对VIIRS数据具有强夜间光探测能力,本文提出了一种卷积运算临界值法对海上油气平台进行提取。首先对2期不同时相的VIIRS数据进行卷积运算,对亮像元进行了增强处理,对背景像元进行弱化处理,从而明确区分疑似目标与背景,然后以0值为分界点,对疑似目标进行提取,最后利用油气平台相对静止的特性,通过邻域分析,实现油气平台的提取。结果表明,本文提出的卷积运算临界值法可以有效地提取油气平台,提取准确率约为85.4%,同时可以有效地减少由于经验阈值对提取结果造成的误差。

DOI

[Li Q, Su F Z, Wang W Y.Research on oil and gas platform extraction technology based on VIIRS data[J]. Journal of Geo-information Science, 2017,19(3):398-406. ]

[20]
高义,王辉,王培涛,等.基于人口普查与多源夜间灯光数据的海岸带人口空间化分析[J].资源科学,2013,35(12):2517-2523.我国海岸带区域是台风、风暴潮、地震海啸和海岸侵蚀等海洋灾害的重灾区,精细空间分辨率的人口数据,能够有效服务海岸带灾害风险管理.本文基于我国第六次人口普查资料、OLS/DMSP和NPP/VIIRS DNB两种夜间灯光数据及Landsat卫星遥感影像,综合利用遥感与地理信息系统理论与技术,进行了我国海岸带人口空间化方法与应用研究.利用建筑物与裸地增强指数法(EBBI)基于Landsat卫星遥感影像提取了我国沿海区县建成区数据,作为人口分布的空间控制因素,以普查人口数与夜间灯光数据回归函数关系为依据,对人口进行空间化处理.反演得到了我国海岸带区县1km×1km和0.5km×0.5km两个空间尺度的人口格网数据.并利用福建省沿海乡镇人口数据对人口空间化结果进行了精度评价.研究结果表明NPP/VIIRS DNB夜间灯光数据适用于人口空间化研究,且其反演精度整体优于基于DMSP/OLS传统夜间灯光数据反演的人口格网模型.通过本文实践,可以发现NPP/VIIRS DNB夜间灯光数据具有实现人口和社会经济数据空间化的巨大潜力.

[Gao Y, Wang H, Wang P T, et al.Analysis on spatial processing of the population for Chinese coastal zones based on census and multiple night light data[J]. Resources Science, 2013,35(12):2517-2523. ]

[21]
杨继瑞,李月起,汪锐.川渝地区:“一带一路”和长江经济带的战略支点[J].经济体制改革,2015(4):58-64.

[Yang J R, Li Y Q, Wang R.Sichuan and Chongqing region: Strategic fulcrum of the Belt and Road initiatives and Yangtze River economic zone[J]. Reform of Economic System,2015(4):58-64. ]

[22]
王炳赟,范广洲,董一平,等.川渝地区气候与物候的变化特征分析[J].地理科学,2011(6):674-681.

[Wang B Y, Fan G Z, Dong Y P, et al.Analysis on the variations of climate and phenology in Sichuan and Chongqing region[J]. Scientia Geographica Sinica, 2011(6):674-681. ]

[23]
Liu J Y, Kuang W H, Zhang Z X, et al.Spatiotemporal characteristics, patterns, and causes of land-use changes in China since the late 1980s[J]. Journal of Geographical Sciences, 2014,24(2):195-210.Land-use/land-cover 变化(LUCC ) 有连接到人和自然相互作用。瓷器 Land-Use/cover 数据集(CLUD ) 从 1980 年代末在 5 年的间隔定期被更新到 2010,与基于 Landsat TMETM+ 图象的标准过程。陆地使用动态区域化方法被建议分析主要陆地使用变换。在国家规模的陆地使用变化的空间与时间的特征,差别,和原因然后被检验。主要调查结果如下被总结。越过中国的陆地使用变化(LUC ) 在最后 20 年(19902010 ) 里在空间、时间的特征显示了一个重要变化。农田变化的区域在南方减少了并且在北方,而是仍然是的全部的区域增加了几乎未改变。回收农田从东北被转移到西北。布满建筑物陆地很快膨胀了,主要在东方被散布,并且逐渐地展开到中央、西方的中国。树林首先减少了,然后增加但是荒芜的区域是反面。草地继续减少。在中国的 LUC 的不同空间模式被发现在之间迟了第 20 世纪并且早第 21 世纪。原版 13 个 LUC 地区在一些地区被边界的变化由 15 个单位代替。包括的这些变化(1 ) 的主要空间特征加速的扩大布满建筑物在 Huang-Huai-Hai 区域,东南的沿海的区域,长江的中流区域,和四川盆登陆;(2 ) 从东北中国和东方内部蒙古在北方转移了陆地开垦到绿洲在西北中国的农业区域;(3 ) 从在到稻的东北中国的喂雨的农田的连续转变回答;并且(4 ) 为在内部蒙古,黄土高原,和西南的多山的区域的南部的农业牧剧的交错群落的格林工程的谷物的有效性。在最后二十年,尽管在北方的气候变化在农田影响了变化,政策规定和经济驱动力仍然是越过中国的 LUC 的主要原因。在第 21 世纪的第一十年期间,在陆地使用模式驾驶了变化的人为的因素从单程的陆地开发转移了强调到开发和保存。动态区域化方法被用来在单位的 zoning 边界,地区的内部特征,和生长和减少的空间17

DOI

[24]
梁友嘉,徐中民.基于LUCC和夜间灯光辐射数据的张掖市甘州区人口空间分布建模[J].冰川冻土,2012,34(4):999-1006.尽管近年来统计数据的生成技术有了很大提高, 但可用的详细人口数据始终难以得到满足.在一些自然-人文要素耦合的建模研究中, 如生态经济集成建模、 环境和健康分析等都需要基于区域尺度的栅格人口分布建模方法.随着GIS和RS技术的发展, 上述建模方法已有较大进步.利用GIS技术, 基于夜间灯光辐射数据和LUCC在象元栅格水平上构建张掖市甘州区人口空间分布. 首先对DMSP夜间灯光辐射数据进行重采样, 通过普通克里金插值获得灯光数据; 然后与LUCC叠加分析, 利用回归分析的方法获取研究区土地利用、 灯光辐射指标和人口统计数据之间的定量关系, 完成空间化.并在乡镇尺度上进行模型检验, 模型总体的调整<i>R</i><sup>2</sup>为0.88, 标准误差为400, 为下一步开展时空变化分析提供支持.

[Liang Y J, Xu Z M.Modeling the spatial distribution of population based on night light radiation data and LUCC: A case study in Ganzhou District, Zhangye municipality[J]. Journal of Glaciology and Geocryology, 2012,34(4):999-1006. ]

文章导航

/