Abstract: Urbanization in China has entered the era of urban regeneration, with an escalating public demand for superior spatial quality. Accordingly, the novel characteristic urbanization of China requires regions to respect social customs and preserve historical features in development. Construction dominated by the demands of local residents can effectively maintain the authenticity of regions. It is crucial for realizing the cultural renaissance of rural areas and guiding the sustainable development of regions. In this study, the ancient town of Shawan in Panyu District, Guangzhou, a semi-urbanized region with a traditional rural social atmosphere, was selected for the case study. The social associations behind the cognitive subjects were disclosed using the mental map, questionnaire survey, and social network analysis. The research route of "image element-cognitive network-influencing mechanism" was established innovatively. Based on Kevin Lynch's theory of city image, five space elements of rural areas, namely markers, nodes, roads, boundaries, and regions, were collected. The functional cognitive networks were established according to the correlations of spatial elements in the cognitive map villagers drew. Besides, establishment logics of villager social networks were analyzed based on the network faction classification in the social network analysis method, thus exploring the social relation features and change mechanisms in typical semi-urbanized regions. This study found that: ① the overall stability of the altars and temples, community, and production cognition network is stronger than that of the cognitive network of the ancestral halls.② The social network behind image cognition in semi-urbanized regions is significantly related to the corresponding material space.③ In  the image cognition process, the interactions among individuals facilitate the production of social relations, and it is closely related to consanguinity, geographical relationship, life, and production needs of villagers.④ Industrial transformation in regions in the process of urbanization influences the overall land use patterns in rural areas, which changes the traditional material spatial structure of the rural regions and affects social relationships of village subjects.⑤ Due to changes in the spatial images and the external migration of core functions in semi-urbanized regions, the cognitive boundaries of villagers from different clans or districts gradually disappear in rural development. The social network presents a trend of multi-relational cooperation. Different integration states are achieved according to different urbanization degrees of village areas. This study introduces the complex network theory into research fields related to city image. Research results provide scientific references to the historical image protection of semi-urbanized regions and novel urbanization development based on human social networks. Some suggestions are facilitating technological innovation in humanism-oriented urban planning and encouraging the local public to participate in urbanization.

Key words: peri-urban area, spatial image, cognitive differentiation, social network analysis, social changes

摘要： 中国特色新型城镇化道路要求地区在发展中尊重社会民俗并保护历史风貌，以本地居民需求为主导的建设有利于实现城镇的人文复兴。以广州市番禺区沙湾古镇为例，运用认知地图圈域法、问卷调查及社会网络分析等方法，创新性地建立 “意象要素-认知网络-影响机制”的研究路径。通过村民认知地图的要素关联来构建功能性意象认知网络，并根据网络派系划分来剖析村民社会网络的构建逻辑，进而解析典型半城镇化地区的社会关系特征与变迁机理。发现意象认知过程中个体间的互动进而产生社会关系；城镇化中的产业转型影响村镇整体土地利用格局，进而影响村民主体的意象认知；半城镇化地区空间意象变化和核心功能外迁使异氏异村村民的认知边界在村镇发展中逐渐消弭，社会网络呈现多元关系协同的趋势。将社会网络理论引入到城市意象的研究领域，为半城镇化地区的历史意象保护与基于人文社会网络的新型城镇化发展提供科学参考和依据。

关键词: 半城镇化地区, 空间意象, 认知分异, 社会网络分析, 社会变迁