基于夜间灯光数据的珠三角地区镇级GDP估算

Township GDP Estimation of the Pearl River Delta Based on the NPP-VIIRS Night-Time Satellite Data

高精度地表GDP分布数据是开展小尺度区域发展相关研究的关键数据,但通常难以获得完整数据序列。文章比较了DMSP-OLS和NPP-VIRS两种夜间灯光数据在小尺度单元GDP估算工作上的适应性,证实NPP-VIIRS数据在镇级GDP估算中有更好的性能。利用修正后的NPP-VIIRS夜间灯光数据建立珠三角地区镇级GDP估算模型,并通过地区差异系数对估计结果进行校正。对2013年镇级GDP估算的实验结果总体精度达到85%。估算结果能够用于填补珠三角地区部分镇级GDP统计数据缺失,为相关研究获取、对比GDP数据提供技术手段。

Though the high resolution data of GDP distribution is difficult to collect, it is very important for small scale regional development research. Night-time satellite images can be used to estimate GDP of administrative unit. In this paper, DMAP-OLS and NPP-VIIRS are used to estimate township GDP of the Pearl River Delta district： DMAP-OLS data are employed to modify NPP-VIIRS data, and then regression model about GDP and NPP-VIIRS Pixel brightness are built to estimate township GDE Regional difference coefficients are employed to adjust the township GDP based on city level GDP data. Experiments show that the estimating accuracy of township GDP value in 2013 a reaches to 85%, and NPP-VIIRS data is valid data resource for township GDP estimation.