基于夜间灯光数据和空间回归模型的城市常住人口格网化方法研究

发布时间：2018-04-19 20:25

本文选题：夜间灯光数据 + 常住人口　；参考：《地球信息科学学报》2017年10期

【摘要】：精确掌握常住人口的数量和分布特征有助于明确社会发展情况、提高人口管理能力。目前人口数据主要以行政区为单元统计,难以表现城市内部的人口分布特征。然而,在城市中,受道路、公共服务设施、城市亮化灯光的影响,利用夜间灯光数据对人口回归,精度降低。如何提高城市常住人口回归结果的精度,值得深入研究。上海是中国的国家中心城市之一,在快速城镇化进程中上海面临巨大人口压力。因此,本文以上海市为研究区,以NPP-VIIRS(National Polar-orbiting Partnership Visible Infrared Imaging Radiometer Suite)夜间灯光数据、乡镇级常住人口统计数据为基础,提取商业和居住区的灯光数据来缓解交通、城市亮化区的影响,提高灯光累计值与常住人口数的相关性(相关系数从0.7032提高至0.8026)。然后,本文用空间回归模型对上海市2013年常住人口进行回归,相对误差为10.57%,并对回归结果进行分乡(镇、街道)修正。实验结果表明,使用空间回归模型对常住人口回归可以取得较高的精度,且格网化结果能够弥补传统统计数据空间分辨率低的缺点,更加详细地刻画常住人口的圈层特征与真实分布情况。
[Abstract]:Accurate understanding of the quantity and distribution of resident population will help to clarify the social development and improve the ability of population management. At present, the population data mainly take the administrative region as the unit statistics, it is difficult to express the urban internal population distribution characteristic. However, in the city, affected by roads, public service facilities, urban lighting, the use of night lighting data to return to the population, the accuracy is reduced. How to improve the precision of urban resident population regression is worth further study. Shanghai is one of the national central cities in China. It faces huge population pressure in the process of rapid urbanization. Therefore, based on the night lighting data of NPP-VIIRS(National Polar-orbiting Partnership Visible Infrared Imaging Radiometer Suite, the light data of commercial and residential areas are extracted to alleviate the impact of traffic and urban lighting areas. The correlation between the light accumulative value and the resident population was improved (the correlation coefficient was increased from 0.7032 to 0.8026). Then, this paper uses the spatial regression model to regression the resident population of Shanghai in 2013, the relative error is 10.57, and the regression result is corrected by the township (town, street). The experimental results show that the spatial regression model can achieve high precision for the permanent population regression, and the grid results can make up for the low spatial resolution of the traditional statistical data. More detailed description of the characteristics of the resident population and the true distribution of the circle.
【作者单位】：江苏省地理信息技术重点实验室;南京大学地理信息科学系;南京市国土资源局;
【基金】：国家自然科学基金项目(41571378) 中国土地勘测规划院外协项目(2016-63-3)
【分类号】：C924.2

【相似文献】