引入路网和建筑物信息的DMSP/OLS数据去饱和方法

A Method to Reduce Saturation of DMSP/OLS Nighttime Light Data for Introducing Road Network and Building Information

由于OLS(operational linescan system)传感器的缺陷,DMSP/OLS数据中的城市中心灯光值存在饱和性。提出了一种基于灯光贡献的综合指数去饱和方法,将路网和建筑物引入到去饱和模型中,并将增强型植被指数(enhanced vegetation index,EVI)作为辅助修正数据对夜间灯光数据去饱和。将该去饱和结果与基于EVI修正的灯光指数(EVI adjusted nighttime light index,EANTLI)从城区内部的地物区分能力、与辐射定标数据的拟合程度、对用电量的估算能力3方面进行比较。结果表明,综合指数在城市内部的细节刻画方面具有明显优势,地物区分能力较高;综合指数与辐射定标数据的整体拟合程度更高,且抽取穿过城区的单行数据拟合其R2最高可达0.928,相比EANTLI可提高0.1;与地区用电量拟合程度同样高于EANTLI,R2可达到0.901。综上,引入路网和建筑物的综合指数能够更好地解决数据饱和问题,且具有更高的可靠性。

A new composite index for efficiently reducing the saturation problem of nighttime light data has been proposed in this study.It introduces the elements of road network and buildings.It assumes that the lighting of road network and buildings contributes more to the DN(digital number)values of the image.The results were analyzed and compared to EANTLI(enhanced vegetation index adjusted nighttime light index)in three aspects:(1)The ability to distinguish objects within city;(2)The degree of fitting with radiation calibration data;(3)The ability to estimate electricity consumption.Experimental results show that composite index has obvious advantages in the details of the interior of the city and the ability to distinguish objects.And the determination coefficient for single row can be up to 0.928,which can be increased by 0.1 compared to EANTLI.And the determination coefficient can reach 0.901 in the estimation of electric power consumption.In conclusion,the composite index can effectively reduce nighttime light data saturation in urban centers and has certain reliability.