Research on Spatial Morphology Analysis of Traditional Villages Based on Machine Learning and Image Segmentation

With the advancement of rural revitalization and urbanization, the protection of traditional village spatial forms has become an important issue to be solved. Although existing studies have used in-terdisciplinary methods to analyze the spatial morphology of traditional buildings, systematic data analysis is still insufficient in the study of traditional villages. For this reason, this study combines machine learning and image segmentation techniques to analyze the spatial morphology of tradi-tional villages in Fulin Village, Quanzhou City, by using UAV tilt photography and digital surface modeling. Through supervised machine learning classification, spatial elements such as the land use types and architectural morphology data are extracted to construct the spatial morphology dataset of the traditional village and analyze it in detail.

随着乡村振兴与城镇化进程的深入推进，传统村落空间形态的保护已成为亟待解决的重要议题。尽管现有研究已采用跨学科方法分析传统建筑的空间形态，但针对传统村落的系统性数据分析仍存在不足。为此，本研究结合机器学习与图像分割技术，采用无人机倾斜摄影（UAV Tilt Photography）与数字表面模型（Digital Surface Modeling, DSM）技术手段，对泉州市福林村的传统村落空间形态展开分析。本研究通过监督式机器学习分类，提取土地利用类型、建筑形态数据等空间要素，构建传统村落空间形态数据集并开展精细化分析。