地理学报 ›› 2015, Vol. 70 ›› Issue (8): 1339-1350.doi: 10.11821/dlxb201508012

• • 上一篇    下一篇

基于EVI指数的DMSP/OLS夜间灯光数据去饱和方法

卓莉1(), 张晓帆1, 郑璟2, 陶海燕1(), 郭宇伯1   

  1. 1. 中山大学地理科学与规划学院 广东省城市化与地理环境空间模拟重点实验室/综合地理信息研究中心,广州 510275
    2. 广东省气候中心,广州 510080
  • 收稿日期:2014-10-27 修回日期:2015-01-31 出版日期:2015-08-20 发布日期:2015-09-24
  • 作者简介:

    作者简介:卓莉, 女, 副教授, 主要从事城市与资源环境遥感、地理模拟、优化算法及灾害风险管理研究。E-mail: zhuoli@mail.sysu.edu.cn

  • 基金资助:
    国家自然科学基金项目(41371499);中山大学柳林教授千人计划科研启动项目

An EVI-based method to reduce saturation of DMSP/OLS nighttime light data

Li ZHUO1(), Xiaofan ZHANG1, Jing ZHENG2, Haiyan TAO1(), Yubo GUO1   

  1. 1. Guangdong Provincial Key Laboratory of Urbanization and Geo-simulation / Center of Integrated Geographic Information Analysis, School of Geography and Planning, Sun Yat-sen University, Guangzhou 510275, China
    2. Guangdong Climate Center, Guangzhou 510080, China
  • Received:2014-10-27 Revised:2015-01-31 Online:2015-08-20 Published:2015-09-24
  • Supported by:
    National Natural Science Foundation of China, No.41371499;The startup fund for Prof.Liu Lin from the One Thousand Talents Program of Sun Yat-sen University

摘要:

DMSP/OLS夜间灯光数据被广泛应用于表征人类活动强度及其生态环境影响的诸多研究中,但因OLS传感器设计的局限,在用电强度较高的城市中心,灯光信号存在明显的饱和,这一不足可能影响到一些基于夜间灯光数据研究成果的可靠性。针对这一问题,NOAA-NGDC研发了辐射定标算法,但因缺乏星上定标系统,算法较为复杂,且受较多条件限制等原因,目前只有部分时期的辐射定标数据产品(RCNTL)。近期有学者提出一种基于植被指数NDVI构建的城市灯光指数VANUI,为灯光数据去饱和研究提供了一个操作简单且结果良好的方法,但该方法在一些城市效果不佳。基于此,本文综合利用夜间灯光与EVI指数信息,通过对VANUI指数构建方法进行改进,建立了一个新的缓解夜间灯光强度饱和的EANTLI指数。为了评价指数的效果,将EANTLI与VANUI从三个方面进行比较:① 区分、识别饱和区内地物的能力;② 与RCNTL的拟合程度;③ 对用电量估算的效果。结果表明:EANTLI在三个方面均表现出优势,在潜在饱和区内对特征地物具有更高的可区分性,与RCNTL的线性相关程度更高,与用电量的相关性相比于NTL、VANUI亦明显提高。因此可以认为EANTLI在指数的设计上较为合理,不仅易于计算,而且能达到较好的缓解灯光强度饱和、凸现城市内部差异的目的,在用于反演城市发展指标时能获得更为准确的结果,因此具有较高的应用价值。

关键词: DMSP/OLS, 夜间灯光, EANTLI, 饱和, EVI指数

Abstract:

DMSP/OLS nighttime light (NTL) data has been widely applied to many studies on anthropogenic activities and their effects on the environment. Due to the limitations of the OLS sensor, NTL data suffers from saturation problem in the core of urban areas, which further influences researches based on nocturnal lights. The radiance calibrated nighttime light (RCNTL) products developed by the National Geophysical Data Center (NGDC) at NOAA partially solved the problem. However, they are only available for a very limited number of years. Recently, a vegetation adjusted NTL urban index (VANUI) has been developed based on the stylized fact that vegetation and urban surfaces are inversely correlated. Despite its simplicity of implementing and effectiveness in increasing variation to NTL data, VANUI does not perform well in some fast growing cities. In this paper, we proposed a new urban index, i.e., the Enhanced Vegetation Index (EVI) adjusted nighttime light index (EANTLI), which combined MODIS EVI with NTL to alleviate the saturation problem of NTL data. In order to evaluate the proposed EANTLI's capability in reducing NTL saturation, we first compared its spatial distributions in potential saturated areas (PSAs) of three metropolitan areas in China with that of the original NTL and VANUI, respectively. Then we randomly selected 30 latitudinal transects across these urban areas to verify EANTLI's similarity to the RCNTL. Finally, we tested EANTLI's effectiveness in assessing electric power consumption of 168 prefecture-level cities in China. Results from these experiments showed that EANTLI significantly increases spatial heterogeneity in the PSAs and effectively alleviates the NTL saturation problem. EANTLI's similarity to RCNTL is consistently higher than that of VANUI in the comparison of latitudinal transects. EANTLI also yields better results in the estimation of electric power consumption. In conclusion, the EANTLI can effectively reduce NTL saturation in urban centers and thus has great potential of wide range applications in the future.

Key words: DMSP/OLS, nighttime light, EANTLI, saturation, EVI