Urban Vacant Land, Underused Space and Stock Renewal

Urban Vacant Land (UVL) is both a resource and a challenge for sustainable urban development. However, large- scale UVL identification remains understudied. This study addresses this gap by proposing an innovative auto - mated UVL identification method utilizing the cutting-edge Segment Anything Model (SAM), applied across all 2446 Natural Cities (NCs) in China. Our findings reveal several distribution patterns. First, the UVL ratio, which refers to the proportion of the UVL area in each NC, follows a log-normal distribution and remains independent of city size. Second, the UVL area adheres to Zipf ’ s law and scaling law, where larger cities tend to have larger UVL areas. Third, we identify five distinct UVL spatial types based on intra-city distribution patterns. In large cities, UVL tends to cluster to form local types, while in smaller cities, they are more dispersed, forming central, pe - ripheral, and scatter types. Forth, regional analysis reveals significant spatial heterogeneity in UVL types across China. Global and peripheral types require special attention, as they present unique challenges due to high UVL ratios and larger average UVL sizes. This study not only advances the methodological framework for UVL identification and providing a comprehensive UVL dataset for China, but also delivers actionable insights for sustainable urban development through the application of AI technology.

第 1 页 (共 1 页) 证书号第8481539号专利公告信息发明专利证书发明名称 ：一种面向城市测度的空地一体大数据融合方法专利权人 ：浙江大学城乡规划设计研究院有限公司浙大启真未来城市科技（杭州）有限公司地址 ：310012 浙江省杭州市西湖区余杭塘路928号3号楼南楼708室发明人 ：张远景;龙瀛;马毅;许雪琳;马学斌专利号 ：ZL 2025 1 1268253.1 授权公告号 ： CN 120744856 B 专利申请日：2025年09月05日授权公告日 ： 2025年11月18日申请日时申请人： 浙江大学城乡规划设计研究院有限公司浙大启真未来城市科技（杭州）有限公司申请日时发明人： 张远景;龙瀛;马毅;许雪琳;马学斌国家知识产权局依照中华人民共和国专利法进行审查，决定授予专利权，并予以公告。 专利权自授权公告之日起生效。专利权有效性及专利权人变更等法律信息以专利登记簿记载为准。 局长申长雨 2025年11月18日 *2025112682531*

中国城市空地的识别与分布规律研究 （申请清华大学工学博士学位论文） 培养单位： 建筑学院学科 ： 城乡规划学研究生： 王新宇指导教师： 龙瀛教授二〇二五年十一月中国城市空地的识别与分布规律研究王新宇 Identification and Spatial Distribution Laws of Urban Vacant Lands in Chinese Cities Dissertation submitted to Tsinghua University in partial fulfillment of the requirement for the degree of Doctor of Philosophy in Urban and Rural Planning by Wang Xinyu Dissertation Supervisor : Professor Long Ying November 2025 学位论文公开评阅人和答辩委员会名单公开评阅人名单党安荣教授清华大学建筑学院答辩委员会名单主席边兰春教授清华大学建筑学院委员汪芳教授北京大学建筑与景观设计学院刘瑜教授北京大学地球与空间科学学院毛其智教授清华大学建筑学院党安荣教授清华大学建筑学院来源副教授清华大学建筑学院龙瀛教授清华大学建筑学院秘书李凤娇助理研究员清华大学建筑学院关于学位论文使用授权的说明本人完全了解清华大学有关保留、使用学位论文的规定，即： 清华大学拥有在著作权法规定范围内学位论文的使用权，其中包括：（1）已获学位的研究生必须按学校规定提交学位论文， 学校可以采用影印、缩印或其他复制手段保存研究生上交的学位论文；（2）为教学和科研目的，学校可以将公开的学位论文作为资料在图书馆、资料室等场所供校内师生阅读，或在校园网上供校内师生浏览部分内容；（3）根据《中华人民共和国学位法》及上级教育主管部门具体要求，向国家图书馆报送相应的学位论文。 本人保证遵守上述规定。 作者签名： 导师签名： 日期： 日期：

• Urban vacant land in 497 shrinking cit - ies on the globe has been identified. • The area and ratio of urban vacant land have increased worldwide from 2016 to 2021. • Spatial distribution patterns of urban vacant land are categorized into six types. • Variation characteristics of urban vacant land reflect socioeconomic differences. • The ratio of urban vacant land presents a phased change in the urbanization process. ARTICLE INFO Editor: Shuqing Zhao

【 目的 】 本文以资源枯竭型城市的低效空间为研究对象 ， 提出一种低效空间的识别方法 , 并以鹤岗市为例 ， 验证了方法的可靠性 ， 构建了城市低效空间数据库 。【 方法 】 通过文献综述 ， 系统性地归纳了资源枯竭型城市所面临的系列空间问题 ， 并结合鹤岗市的实际情况 ， 选择采矿塌陷区 、 城市空地 、 失序空间和废弃建筑进行研究 。 在现有数据基础上 ， 创新性地引入了基于深度学习模型的自动检测技术 ， 基于城市遥感图片 、 城市街景图片等图片数据源 , 完成了针对 4 类低效空间的识别 。【 结果 】 本文采用 DeepLab V 3 模型和 SegNet 模型生成了鹤岗市低效空间数据集 ， 并通过实地调研对识别结果进行了完善 。 研究形成了鹤岗市低效城市空间数据库 ， 并分析了采矿塌陷区 、 城市空地 、 失序空间和废弃建筑在城市中的分布情况 。【 结论 】 鹤岗市的实际应用证实了研究方法能够高效 、 快速 、 准确地识别城市尺度的低效空间 ， 为资源枯竭型城市的低效空间识别提供了有效的技术支持 。 此外 ， 本文提出的研究方法在识别对象的定义 、 技术细节等层面依然存在改善空间 ， 亟待后续研究完善 。

The number of abandoned buildings in shrinking cities is increasing sharply, posing environment risks, threatening the safety and health of residents, affecting the real estate market, and burdening government finance. Abandoned building detection provides fundamental information for refined urban management, real estate transactions and government decision-making. However, emerging sources of data, such as satellite imagery and commercial street views, are insufficient to timely collect this fine-scale data, lacking large-scale and fine-grained detection method. Therefore, in this research, we aim to define the connotation and identification criteria of abandoned buildings, develop an effective deep learning method based on image segmentation, and detect individual abandoned buildings from large-scale mobile sensing images (MSIs) with high accuracy. The study conducted a mobile sensing campaign in a shrinking city in Northeast China, collecting 11,359 street-level images of 126.2 km of urban roads. The accuracy of the deep learning detection method was 83.8%. The study compared with the detection of commercial street view images (latest in 2015) and analyzed the dynamic changes of abandoned buildings. From 2015 to 2021, the number of abandoned buildings in the case city decreased from 102 to 50 and became more concentrated in the old city area. Our study demonstrates the feasibility of MSIs in detecting abandoned buildings and shows the enor- mous potential to timely detect abandoned buildings in large spatial ranges.

Urban vacant land is a growing issue worldwide. However, most of the existing research on urban vacant land has focused on small-scale city areas, while few studies have focused on large-scale national areas. Large-scale identification of urban vacant land is hindered by the disadvantage of high cost and high variability when using the conventional manual identification method. Criteria inconsistency in cross-domain identification is also a major challenge. To address these problems, we propose a large-scale automatic identification framework of urban vacant land based on semantic segmentation of high-resolution remote sensing images and select 36 major cities in China as study areas. The framework utilizes deep learning techniques to realize automatic identification and introduces the city stratification method to address the challenge of identification criteria inconsistency. The results of the case study on 36 major Chinese cities indicate two major conclusions. First, the proposed frame - work of vacant land identification can achieve over 90 percent accuracy of the level of professional auditors with much higher result stability and approximately 15 times higher efficiency compared to the manual identification method. Second, the framework has strong robustness and can maintain high performance in various cities. With the above advantages, the proposed framework provides a practical approach to large-scale vacant land iden - tification in various countries and regions worldwide, which is of great significance for the academic develop - ment of urban vacant land and future urban development.