精细尺度下的城市人口空间分布研究可为城市人口疏导制定科学合理的方案，对于解决快速人口城镇化带来的人口空间聚集、区域发展失衡、资源分配失调等问题具有重要意义。以重庆市中心城区为例，提出了一种基于建成环境的城市人口职住分布空间化方法：首先基于多源地理数据获取建成环境变量；然后通过计算建成环境变量与职住人口的灰色关联度，实现强相关建成环境变量的筛选；最后采用随机森林模型对所筛选变量进行训练，构建职住人口空间化模型。结果表明，本方法对于职住人口空间化拟合精度较高，可以有效地运用于模拟城市职住人口空间分布，从而精准地掌握职住人口空间分布情况。

A method was proposed for spatial distribution analysis of urban population in the central urban area of Chongqing city, aiming to formulate scientific and rational schemes for population dispersion. This method was based on the built environment and involves the following steps: firstly, acquiring built environment variables using multi-source geographical data; secondly, calculating the grey correlation between built environment variables and residential population, thereby filtering strongly correlated variables; and finally, training a random forest model on the selected variables to construct a spatial model of residential population distribution. The results indicate that this method achieves a high fitting accuracy for residential population spatialization, effectively to simulate urban residential population distribution and accurately understand the spatial distribution of residential population.