基于遥感影像的城市人口密度模型

分析了常用的城市人口遥感估算方法,指出了其特点和不足,在土地利用密度法的基础上作了改进。土地利用密度法的关键是抽样区的代表性,抽样区选择的不当,就会给人口计算带来较大的误差。实际上,抽样区的选择有较大的难度,而且实地区域和抽样区相比总会有一定的差异,所以人口估计误差是不可避免的。作者试图避开抽样区的选择而估计出各种居住类型的人口密度,提出的方法不需要实地抽查人口密度,减少了估算的工作量,建立的数学模型没有抽样的随机性误差影响,结果从整体上达到了最优。

Modeling Urban Population Density with Remote Sensing Imagery

The population in the study region is the sum of the area of each habitation type multiplying its respective sampled population density. The population density of each habitation type is estimated. In some experimental region, there is mathematical relation among the area, the population density and its total population of each habitation type. The population density of each habitation type can be estimated via the mathematical relation. A new land use density method is proposed based on least square principle. The main idea is: First, habitation type is defined in the study region; and the boundaries of all the habitation types are lined out based on remote sensing imagery and the area of each habitation type are calculated; then mathematical models according to the population data of every sub-region are established; the best population density estimation of each habitation type is calculated with the least square principle when the number of sub-regions is more than the number of habitation types. The population estimation of any region can be calculated since the population density of each habitation type is known as well. The method need not sample the population density of each habitation type. The estimation workload of population density of each habitation type is low. The mathematical models are not influenced by random error of samples.