Simulation Of China's Socioeconomic Indicator Based On Long-time Series Nighttime Light Remote Sensing

At present,the background noise problem of the new generation NPP/VIIRS night light remote sensing image is the main reason for restricting the wide application of this data.In view of the great potential of NPP/VIIRS nighttime light images to effectively estimate socioeconomic indicators at large scales,this study proposed to simulate the spatial-temporal dynamics of a certain socioeconomic indicator in Mainland China by using NPP/VIIRS data from 2012 to 2017.Firstly,based on the analysis of the source and distribution of noise,the original NPP/VIIRS nighttime light data was preprocessed by a variety of methods,and a noise removal method suitable for NPP/VIIRS long-time series data was proposed.Secondly,to further test the noise removal effect as well as study the advantages of NPP/VIIRS data in the simulation of socioeconomic indicators,regression analysis comparison for a number of socioeconomic indicators with NPP/VIIRS data as well as DMSP/OLS data were performed,and key socioeconomic indicators were selected from them.Finally,after examining the reliability of using the corrected NPP/VIIRS long-time series data to estimate the key indicator,spatial models were constructed to simulate its spatial changes based on pixels,thus realizing the monitoring of the spatial-temporal dynamics of the relevant socioeconomic in Mainland China.The study indicates that:1)Three different preprocessing methods used in this article could effectively reduce the negative impacts from the background noise of the original NPP/VIIRS data and provide more reliable data sets for the estimation of socioeconomic indicators,and the Noise Threshold Method was more suitable for noise removal processing of NPP/VIIRS long-time series data;2)The corrected NPP/VIIRS data was more advantageous than DMSP/OLS data in the simulation of Total Population,Urban Population,Gross Domestic Product,Secondary Industry,Tertiary Industry,Passenger Traffic,Freight Traffic,Electric Power Consumption,Urban Construction Land Area,Urban Residential Land Area and Urban Road Area,and Electric Power Consumption was the key socioeconomic indicator closely related to nighttime lights;3)The model methods based on the corrected NPP/VIIRS long-time series nighttime light datas and Electric Power Consumption statistics in this paper had high accuracy in both estimation and spatial simulation of Electric Power Consumption;4)The increase rate and spatial expansion trend of Electric Power Consumption in Mainland China from 2012 to 2017 were relatively low,centered on Electric Power Consumption high value regions of provincial capitals,high value area continues to expand and gradually spread to surrounding cities.Meanwhile,spatial connectivity of adjacent high value regions was gradually increasing,and the overall spatial distribution of Electric Power Consumption showed a pattern of high in coastal,low in inland,high in east and low in west.