中文摘要：

为解决以行政区为单位的人口数据空间表示法无法准确反映人口实际分布的问题，文章以广州市天河区为例，利用高分辨率遥感影像数据，采用面向对象法对建筑进行提取和分类，并结合广州市土地利用现状图对分类结果进行验证和修订。在此基础上，利用土地利用密度法估算研究区各街道居住建筑的人口密度，并采用格网法进一步构建城市人口空间分布模型模拟城市居住人口的空间分布。结果表明，广州市天河区的人口主要集中在广园快速公路以南的带状区域内，中心商业区和工业密集区的人口密度最大，主干道附近区域的人口密度显著大于其他区域；距离城市中心较远的新塘、凤凰、长兴、龙洞等街道的人口密度整体偏低，模拟结果与研究区的实际情况相符。文章采用的数据可获得性强，操作方法简单，估算精确度高；研究提高了建筑解译和分类的效率，真实地反映了城市人口的空间分布特征，建模方法和研究思路具有一定推广意义。

英文摘要：

High-resolution remote sensing data of Tianhe District of Guangzhou were used to extract the information of urban buildings by object-oriented method. The extracted information was classified into different kinds with reference to land use map of Guangzhou. On that basis the data of the residential buildings were used to estimate population density. Then an urban population distribution model was created. The operation of other existing population density models needs much time to extract building information from remote sensing images by manual work and can only be used for small scale of settlement. In this study, the efficiency of image classification and extraction of building information was greatly improved, and the accuracy of the model was higher than that in other existing researches. The simulation result indicates that the population density model can capture most typical features and interior differentiation of urban population distribution. The model and method in this study have advantages in accuracy, simplicity, availability and convenience for use.