Night light data DMSP/OLS has been confirmed to be related to economic indicators,but the original data has many prob-lems such as multi-sensor deviation and invalid lighting,and cannot be directly used in the study of economic indicators. Therefore,this paper first uses the dichotomy to find the threshold generation mask to extract the original data,excludes the influence of the cells below the threshold,uses the averaging method to solve the deviation caused by different sensors in the same year,constructs the night light indexes,establishes a variety of regression models with the GDP data of Beijing over the years,and finally uses the linear regres-sion model with the smallest relative error to invert the GDP of Beijing,Its correlation coefficient reaches 0.986 and the inversion effect is better.
万方数据
维普
