基于DMSP/OLS夜间灯光影像的GDP预测模型研究

Research on GDP Prediction Model Based on DMSP/OLS Night-time Light Image

针对经济数据统计烦琐、滞后且不准确等问题,利用夜间灯光影像DMSP/OLS进行江苏各地级市GDP预测研究。对影像二值化处理,确定夜间灯光影像阈值,结合GDP数据构建多种常规回归模型。提出光亮面积指数参与建模的方法,预测经济数据并进行验证。实验结果显示,利用光亮面积指数法进行经济预测,可以提高预测的可靠性与科学性,具有可行性。

In the view of the problem of the tedious,lagging and inaccurate statistics of economic data,the night light image DMSP/OLS was used in the study to predict GDP in prefecture-level cities of Jiangsu Province.The image finalization processing was used to determine the threshold of night light image and a variety of conventional regression models were constructed based on GDP data.A regression model construction method was proposed based on light area index and GDP data and the predicted value of economic data was obtained and verified.The study shows that the use of the light area index method for economic prediction can improve the reliability,scientific nature and feasibility of the prediction.