【摘要】 历史时期聚落时空演化对于认识过去环境变迁、土地利用和人地关系变化具有重要意义。中国历史文献蕴含丰富的聚落信息，是开展距今百年以上聚落演化研究的重要代用资料。历史文献一般用地名记录聚落的空间位置，然而现有的历史地名空间数据库难以实现聚落等微观地名的自动定位，制约了历史聚落空间重建及时空演化研究的推进。本文聚焦历史聚落地名的自动匹配方法，以清代晚期的苏州府为研究区，选取地方志记录的11340个聚落地名及其基层区划资料，结合现代地名匹配技术构造古今地名匹配概念模型，归纳地名演变特征，构建地名匹配规则，提出基于相似度的聚落地名分组匹配算法，完成研究区历史聚落的空间重建。在此基础上对重建结果的空间分布和城乡关系进行分析。结果显示，基于地名自动匹配方法重建的聚落数量约占文献记录总数的98%，空间分辨率约为1.4 km×1.4 km，自动匹配结果与人工考证结果的一致度约96%；清代晚期研究区聚落分布呈集聚特征，以苏州城西郊为核心形成聚落网络等级结构；网络内部城乡联系散漫。研究结果改进了利用历史文献进行聚落空间重建的技术手段，深化了对长期聚落演变阶段特征及其影响因素的科学认识。更多还原

【Abstract】 The spatiotemporal evolution of settlements during historical periods is of great significance for a better understanding of past environment changes, land use, and the dynamics of human-land relationships. Chinese historical documents contain abundant information on settlements, serving as an important proxy for the study on the evolution settlements over the past century. Historical documents typically employ place names to record the spatial location of settlements, but the existing historical place names spatial databases often fall short in automating the precise localization of micro-scale place names, such as settlements, which in turn impedes the advancement of research on the spatial reconstruction and spatiotemporal evolution of these historical settlements. This study focuses on the automated matching method of historical settlement place names, utilizing Suzhou Prefecture during the late Qing Dynasty(1820-1911) as the research area. It selected 11340 settlement place names and their associated primary-level division data extracted from local chronicles, and combined contemporary place names matching techniques to construct a conceptual model for correlating ancient and modern place names. The study identifies characteristics of place names evolution, establishes rules for place names matching, and proposes an algorithm for the aggregation and matching of settlement names predicated on similarity metrics, thereby accomplish the spatial reconstruction of historical settlements within the research area. On this basis, the spatial distribution and urban-rural relationships of the reconstruction results were analyzed. The results indicated that approximately 98% of the settlements, relative to the total number documented in the literature, were reconstructed using an automated place names matching method. The spatial resolution of the reconstruction was approximately 1.4 km × 1.4 km, and the consistency between the automatic matching outcomes and the manually verified results was about 96%. During the late Qing Dynasty, the distribution of settlements in the study area exhibited agglomerative characteristics, with a hierarchical structure of settlement networks centered on the western suburbs of Suzhou city. However, urban-rural connections within the network were sparse. The findings from this study have enhanced the technical methods for the spatial reconstruction of historical settlement space and deepened the scientific understanding of the characteristics and influencing factors of long-term settlement evolution.更多还原