城市街区形态的深度解析与智能分类探索——以杭州、南京为例  

Exploration of Intelligent Classification and In-depth Analysis of Urban Block Morphology:A Case Study of Hangzhou and Nanjing

作　　者：张豪 宣蔚[1,2] Zhang Hao;Xuan Wei

机构地区：[1]合肥工业大学建筑与艺术学院 [2]合肥工业大学安徽省城市风貌与空间环境更新技术工程研究中心

出　　处：《华中建筑》2025年第3期109-114,共6页Huazhong Architecture

基　　金：国家自然科学基金项目(编号:52008143);安徽省哲学社会科学规划项目(编号:AHSKY2021D74)。

摘　　要：当代城市发展日趋复杂,对于大尺度、拼贴化的城市形态,传统的解析方法因数据的扁平化,分析结果不能真实地反映城市空间形态特征。该研究立足量化分析方法,在经典理论的基础上融合多源数据,尝试构建以街区为研究单元的城市形态定量解析。以南京、杭州两座城市的主城区为例,对不同街区相同维度进行对比分析,总结两个城市的形态特征与异同。同时通过无监督机器学习的方法对各街区解析特征指标进行分类,并探讨城市形态分类结果的合理性。Contemporary urban development is becoming more and more complex,and for large-scale and collage urban forms,the traditional analysis methods cannot truly reflect the characteristics of urban spatial morphology due to the flattening of data.Based on the quantitative analysis method,we integrate multi-source data on the basis of classical theory,and try to construct a quantitative analysis of urban morphology with blocks as the research unit.Taking the main urban areas of Nanjing and Hangzhou as examples,we compare and analyze the same dimensions of different neighborhoods,and summarize the morphological characteristics and similarities and differences of the two cities.At the same time,the unsupervised machine learning method is used to classify the analytical feature indicators of each block,and the rationality of the urban form classification results is discussed.

关 键 词：城市形态量化解析 无监督机器学习 城市形态学 大尺度城市模型 

分 类 号：TU984.2[建筑科学—城市规划与设计]