基于波士顿矩阵的夜光遥感电力消费估算方法

Estimation Method of Nighttime Light Images’ Electric Power Consumption Based on the Boston Matrix

针对夜光遥感影像估算社会经济参数在局部区域精度不足的问题,提出了基于波士顿矩阵的灯光指数法。该方法在传统灯光指数法的基础上引入波士顿矩阵,将社会经济参量中的电力消费量作为研究对象,以2015年可见光红外成像辐射仪(visible infrared imaging radiometer,VIIRS)年产品为主要数据源,选择中国大陆区域182个城市作为研究区域,利用波士顿矩阵将全部城市分为4类:明星城市、现金牛城市、问题城市与瘦狗城市。结果显示,基于波士顿矩阵的灯光指数法的平均相对误差为34.04%,小于传统灯光指数法的平均相对误差41.69%;且分类后的各类城市,除明星城市外,基于波士顿矩阵的灯光指数法的平均相对误差均小于传统灯光指数法,而两种方法对于明星城市电力消费估算的结果相差不大;将估算结果按照相对误差的大小分为高精度、中等精度与低精度3种,基于波士顿矩阵的灯光指数法高精度与中等精度的估算结果比例均高于传统灯光指数法,而低精度结果的比例则低于传统灯光指数法。由此可见,基于波士顿矩阵法的灯光指数法对于电力消费量的估算结果优于传统灯光指数法。

Aiming at the problem of inadequate accuracy of social and economic parameter estimation in local area by nighttime light images, a new light index method based on the Boston matrix is proposed. Based on the traditional light index method, this paper introduces the Boston matrix and establishes the linear model after classifying all the study areas. We choose the electric power consumption of social and economic parameters as the research object, with 2015 VIIRS product as the main data source, select a total of 182 prefecture-level cities in China's Mainland from the 17 provinces, autonomous regions and municipalities as research areas. By using the Boston matrix,all cities are divided into four categories: star city, cash cow city, problem city and thin-dog city, respectively.The results show that the average relative error of light index method based on the Boston matrix is 34.04%, which is less than 41.69% of the traditional light index method. In addition, the average relative error of the light index method based on the Boston matrix is smaller than that of the traditional light index method, except for the star cities.According to the relative error, the estimation results are divided into three types: high precision, medium precision and low precision. The proportion of the high and medium precision estimation results based on the Boston matrix is higher than that of the traditional light index method, while the proportion of the low-precision results is lower than that of the traditional light index method.It can be seen that the light index method based on the Boston matrix me-thod is superior to the traditional light index method in the estimation of electric power consumption.