DMSP/OLS数据去饱和处理与省域生产总值反演  

DMSP/OLS Data Desaturation and Chinese Gross Regional Product Regression

作　　者：王毓乾[1] 孔轩 钱宽 程朋根[1] WANG Yuqian;KONG Xuan;QIAN Kuan;CHENG Penggen(Faculty of Geomatics,East China University of Technology,Nanchang 330013,China)

机构地区：[1]东华理工大学测绘工程学院

出　　处：《遥感信息》2019年第4期16-21,共6页Remote Sensing Information

基　　金：科技部科技计划项目(2017YFB0503700);国家自然科学基金项目(41861052);东华理工大学教学改革研究课题(DHJG-07-34);东华理工大学博士科研启动基金项目(DHBK2015309)

摘　　要：针对DMSP/OLS夜光遥感影像像元的亮度值在城市中心区域普遍存在的饱和问题,提出了一种新的去饱和纠正方法,并利用纠正后的影像进行省域生产总值(gross regional product,GRP)反演。由于植被指数与夜光强度存在负相关关系,首先利用增强型植被指数(enhanced vegetation index,EVI)对饱和地区的像元值进行去饱和纠正,非饱和地区的像元值保持不变,然后利用纠正后的夜光强度数据与省域GRP进行线性回归分析,建立GRP的反演模型。本文利用MODIS的EVI产品数据对中国大陆(不包括港澳台地区)2001年至2013年间的DMSP/OLS数据进行了去饱和处理,然后建立省域GRP线性回归模型。实验表明,该算法有效缓解了DMSP/OLS数据的饱和问题,提高了GRP的预测精度。A novel DMSP/OLS night light image desaturation correction method is proposed in this paper to handle the problem of pixel intensity saturation of DMSP/OLS images commonly occurred in the cores of cities.Then the corrected images are used in the inversion of Chinese provincial gross regional product(GRP).As vegetation index has negative relationship with night light intensity,the pixel DN values in saturated region are corrected by enhanced vegetation index(EVI)data,while other pixels remain unchanged.Then a GRP inversion model is established by a linear regression analysis between corrected night light intensity and GRP of Chinese provinces.In experiment part,EVI product data from MODIS are introduced to correct DMSP/OLS data from 2001 to 2013 of main land China(excluding HongKong,Macao and Taiwan regions).Then provincial GRP linear regression model is established with corresponding GRP data.The comparative experiments show that the proposed method could efficiently alleviate the saturation problem of DMSP/OLS data and promote the accuracy of GRP prediction model.

关 键 词：夜光遥感 DMSP/OLS EVI指数 去饱和 GRP预测 线性回归分析 

分 类 号：P407.8[天文地球—大气科学及气象学]