Spatialization Of Population Density Based On Night Time Light Images And The Correlation Analysis Between PM2.5 And People Density

With the acceleration of urbanization process,the population of urban areas is becoming more and more concentrated,and the impact of human activities on the transformation of nature is also increasing.Night lighting,as a new data source for monitoring human activities,is often used to estimate socio-economic parameters,assess light pollution and major events,and dynamically monitor urbanization and fishery monitoring.However,there are some problems in the existing night lighting data,such as saturation of urban central pixels,lack of radiation calibration on stars,lack of comparability of image data,and inconsistency in the scale of multi-source night lighting image radiation brightness.These problems may lead to the time discontinuity of night lighting image,and also affect the accuracy of data.In order to solve the above problems,a method of extracting invariant target region based on linear fitting is proposed.The night light images from DMSP and VIIRS data sources are mutually corrected.On this basis,combined with the census data?provinces,cities,counties?,land use,DEM elevation data,vegetation index and other data of the China-Pakistan Economic Corridor,the stochastic forest model is adopted to realize the spatialization of population density in the whole area of the China-Pakistan Economic Corridor.Finally,this paper preliminarily explores the impact of human activities on the spatial distribution of PM_2.5 concentration in the China-Brazil economic corridor by using correlation analysis method.The main work and conclusions of this paper include:?1?In order to solve the problem of pixel saturation and discontinuity of time series image in night remote sensing image center area,this paper uses linear fitting method to extract the invariant region between images,and establishes the linear model between reference image and image to be corrected in the invariant target area,so as to realize the mutual correction between DMSP image,DMSP and VIIRS data.The results show that the goodness of fit of the two calibration models is above 0.78.The correlation between the total gray level of the corrected DMSP image and GDP and population data is significantly improved?GDP:R²=0.7689;population:R²=0.9033?,and the standardized difference index is significantly reduced.After mutual correction,the VIIRS image is more consistent with DMSP in radiation brightness,spatial and temporal distribution,and spatial details are more prominent.The consistency of the images was enhanced.?2?In this paper,the population density in the lighting area of China-Brazil economic corridor is simulated by using the corrected night light data?DMSP and VIIRS?,land use,DEM,vegetation index and so on,and its change characteristics are analyzed from two dimensions of time and space.The results show that the overall accuracy of the simulation is 82.15%,slightly lower than that of GPW?89.45%?and World Pop?89.75%?.From the point of view of the importance of variable factors,land use and night light index are important indicators of population spatial distribution.Spatially,the middle and high value areas of population density in China-Pakistan economic corridor are mainly distributed in the arc zone formed by Peshawar,Islamabad,Lahore,Fesalabad,Muertan,Sukur,Karachi and other cities.From the time point of view,the population density of the whole study area increased steadily from 2000 to 2015,and showed the difference of growth rate in different time periods and regions.?3?The correlation between night light intensity and population density representing human activities and PM_2.5 concentration in the study area was preliminarily explored by correlation analysis method.The results show that the areas with high population density and high human activity intensity revealed by the corrected stable light images have a good spatial correspondence with the areas with high PM_2.5 concentration.In terms of time,there is a good correlation and correspondence between the average night light intensity and population density,but the response of PM_2.5 concentration to population density growth has a certain lag.From the correlation point of view,the regional population density,night light intensity mean and PM_2.5 concentration mean were significantly positively correlated,and the correlation between population density and PM_2.5 mean was lower than the correlation between the mean of night light intensity and PM_2.5 concentration.