Spatial Disorder, Evaluation and Renewal Strategies

Physical disorder can cause social disorder and urban decline. The emergence of virtual auditing via online street-view images provides a brand-new way to study physical disorder in the real world. However, most precedent studies focused on cities in developed nations, leaving cities in the developing world largely unstudied. This study aims at addressing this lacuna by taking the City of Qiqihar (Heilongjiang, China) as an example, to decipher its physical disorder status and related factors. By geo-sampling, we extracted 4852 street-view images from Tencent Maps. After training, auditors developed a checklist with 13 disorder items and independently audited the 4852 images to identify disorder items and perceive disorder of each image. We found that the number of disorder items in a street-view image is strongly correlated with the chance of its being perceived as disorder by auditors. In addition, seven out of the 13 disorder items are more strongly associated with the perceived disorder when compared with other disorder items. This study demonstrates that virtual auditing via street-view images is feasible for researching physical disorder and related issues in Chinese cities. It also contributes to the refinement of the traditional “broken windows theory”.

Annals of the American Association of Geographers ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/raag21 Measuring Temporal Evolution of Nationwide Urban Physical Disorder: An Approach Combining Time-Series Street View Imagery with Deep Learning Yue Ma, Yan Li & Ying Long To cite this article: Yue Ma, Yan Li & Ying Long (07 Mar 2025): Measuring Temporal Evolution of Nationwide Urban Physical Disorder: An Approach Combining Time-Series Street View Imagery with Deep Learning, Annals of the American Association of Geographers, DOI: 10.1080/24694452.2025.2467330 To link to this article: https://doi.org/10.1080/24694452.2025.2467330 Published online: 07 Mar 2025. Submit your article to this journal Article views: 28 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=raag21 Measuring Temporal Evolution of Nationwide Urban Physical Disorder: An Approach Combining Time-Series Street View Imagery with Deep Learning Yue Ma, a Yan Li, a,b and Ying Long c aSchool of Architecture, Tsinghua University, China; bSchool of Land Science and Technology, China University of Geosciences, China;

Neighborhood physical disorder (NPD) has been theorized to be associated with health outcomes at an individual level, with inconsistent evidence across studies. To address these discrepancies, a systematic review is imperative to unravel the discrepancies present in the literature and offer an evidence-based assessment of the association between NPD and individual health. This systematic review analyzed 50 studies that adhered to the inclusion criteria out of the 4185 articles retrieved. The majority of the studies included in this review were conducted in the United States and employed cross-sectional designs. The measurement of NPD predominantly relied on traditional methods, such as interview or questionnaire, and most studies focus on individuals’ physical health and health-related behaviors, with relatively less attention paid to mental health, social health and self-rated health. Findings from the systematic review revealed that 20 cross-sectional studies showed a correlation, while 17 did not. Additionally, 10 cohort studies identified an adverse effect of NPD on health outcomes, whereas 8 cohort studies and 1 case–control study reported null findings. Subgroup analysis showed that the diversity of NPD types, variations in the indicator systems employed, and disparities in the definition of the neighborhood contributed to the heterogeneity of findings.

Protocol Protocol for assessing neighborhood physical disorder using the YOLOv8 deep learning model Neighborhood physical disorder (PD), characterized by disruptions and irregularities in spatial elements, is associated with negative economic, social, and public health outcomes. Here, we present a protocol to quantitatively assess PD utilizing a range of metrics. We describe steps for collecting street views, constructing detection models using the YOLOv8 deep learning model, calculating PD scores, and quantifying changes in PD across streets and cites. This protocol serves as a methodological foundation for assessing PD in different countries and regions. Publisher’s note: Undertaking any experimental pro tocol requires adherence to local institutional guidelines for laboratory safety and ethics. Yan Li, Yue Ma, Ying Long yanli427@hotmail.com (Y.L.) ylong@tsinghua.edu.cn (Y.L.) Highlights Construct a dataset with commercial and customized street view images over time Implement deep learning YOLOv8 model to standardize disorder calculations Calculate disorder scores at city, street, and specific location granularity Li et al., STAR Protocols 5, 102778 March 15, 2024 ª 2023 The Authors.

In the process of advancing urban renewal and stock planning through spatial assessment and problem diagnosis, the quality and accuracy of data directly impact the outcomes of spatial evaluations and the direction of planning and design, thus making the methods of data acquisition and utilization crucial. Traditional urban spatial assessment methods, such as surveys and interviews, are costly and depend heavily on subjective judgments; meanwhile, the use of open big data often lacks in terms of applicability and timeliness. This paper focuses on exploring a multi-source data acquisition approach that integrates emerging proactive sensing technologies and big data collection with conventional urban renewal survey methods. Taking the central urban area renewal project in Tancheng County, Shandong Province as an example, to support the formulation and execution of related urban renewal strategies and practices, it combines new, high-precision, and extensive coverage data sources like proactive sensing street views and open remote sensing images. By using artificial intelligence methods, it identifies the current characteristics of urban spaces, constructs a spatial assessment indicator system, and diagnoses urban spatial issues, while concurrently employing traditional survey methods to understand current conditions and renewal demands. This paper establishes a systematic research pathway of “data collection — database construction — spatial assessment — strategy development for renewal”, providing a methodological reference for the conduct of urban spatial evaluation and renewal practices. 关键词 ：主动感知技术 ； 城市大数据 ； 空间评价 ； 城市更新 ； 技术方法

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=raag21 Annals of the American Association of Geographers ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/raag21 Measuring Physical Disorder in Urban Street Spaces: A Large-Scale Analysis Using Street View Images and Deep Learning Jingjia Chen, Long Chen, Yan Li, Wenjia Zhang & Ying Long To cite this article: Jingjia Chen, Long Chen, Yan Li, Wenjia Zhang & Ying Long (2022): Measuring Physical Disorder in Urban Street Spaces: A Large-Scale Analysis Using Street View Images and Deep Learning, Annals of the American Association of Geographers, DOI: 10.1080/24694452.2022.2114417 To link to this article: https://doi.org/10.1080/24694452.2022.2114417 View supplementary material Published online: 14 Oct 2022. Submit your article to this journal View related articles View Crossmark data Measuring Physical Disorder in Urban Street Spaces: A Large-Scale Analysis Using Street View Images and Deep Learning Jingjia Chen, /C3 Long Chen, † Yan Li, /C3 Wenjia Zhang, ‡ and Ying Long § /C3 School of Architecture, Tsinghua University, China †Beijing Key Laboratory of Traffic Engineering, Beijing University of Technology, China, and School of Architecture, Tsinghua University, China ‡Peking University Shenzhen Graduate School, China §School of Architecture and Hang Lung Center for Real Estate, Key Laboratory of Eco Planning & Green Building, Ministry...

To date, most studies have assessed individual exposure to neighborhood physical disorder (NPD) through the static residence-based approach, which ignores elements of human mobility and may lead to inaccurate esti - mates. This study assessed individual exposure to neighborhood physical disorder through the mobility-based approach using wearable cameras. The use of this approach allowed us to leverage innovative tools to accu - rately assess exposure to NPD in individuals ’ activities in space-time. We assessed the volunteers ’ exposure to neighborhood physical disorder by manually auditing pictures taken by wearable cameras on an online browser- based assessment platform. The results illustrated that wearable cameras can clearly capture the exposure while volunteers were engaged in travel behaviors. We also compared the proposed approach (mobility-based, using wearable cameras to take photos) with other approaches (with consideration of travel behaviors to varying degrees, using street view images) to demonstrate that wearable cameras can record individual exposure to neighborhood physical disorder accurately and conveniently, and the assessment results might be significantly different from those obtained by other approaches. Thus, the proposed approach is of great significance.

空间失序视角下的城市街道空间品质测度研究 (申请清华大学城市规划硕士专业学位论文) 培养单位 ： 建筑学院申请人 ： 陈婧佳指导教师 ： 龙瀛副教授二○二一年六月空间失序视角下的城市街道空间品质测度研究陈婧佳 Research on the Quality Measurement of Urban Street Space from the Perspective of Spatial Disorder Thesis Submitted to Tsinghua University in partial fulfillment of the requirement for the professional degree of
Master of Urban Planning
by Chen Jingjia
Thesis Supervisor : Long Ying, Associate Professor June, 2021



<("=C46? 321
@H 
°

Ů
ÁɃȦ
ˋÛŏť
Ϋɘ̟
қ͡Ƙ̺Ƣ
ÔEļƠϴѯ̳ȇɬѯă
7D5. 
+ƥ
dž

ǫ
ɃȦ
ˋÛŏť
ŜĒ
°

Ů
ÁɃȦ
ˋÛŏť

°l˦
ÁɃȦ
ˋÛŏť

ң

˚
ÁɃȦ
ˋÛŏť

ʑǣǣ
ÁɃȦ
ˋÛŏť

džѲƓ
ɃȦ͡қ͡Ƙ̺Ƣ
ÔEˋÛĆοϊ®Ϟϓ̤̼Ѱɬѯă
̴7
义ɨ
àň̤̼˽
ˋÛŏť

时代建筑 Ｔｉ ｍ ｅ＋ Ａ ｒ ｃ ｈ ｉ ｔ ｅｃ ｔ ｕ ｒ ｅ ２ ０２ １ ／ １ ４ ５ 行客观且可靠的比较 ， 使得与人密切相关的城市公共空间难以被大规模量化和测度 。 随着城市研究可获取的新数据环境的形成 ， 基于街景图片的非现场建成环境审计 （ “ 虚拟审计 ” ， ｖ ｉ ｒｔ ｕ ａ ｌ ａ ｕ ｄ ｉ ｔ ） 以及其衍生开发的在线工具平台开始应用在建成环境审计及测度研究中 。 国外的谷歌街景 、 国内的腾讯街景和百度街景等数据平台 ， 提供了可获取的并且覆盖大部分城市主城区街道的高精度街景图片 。 一方面 ． 基于街景图片开展的城市研究通常包括了三类 ： 对图片元数据的挖掘如拍照地点 、 时间等 ， 对图片文本标签的挖掘 ， 以及对图片内容本身的挖掘 ［ ６１ 。 前两类研究主要用于城市形态分析和人群的时空行为分析而对图片内容信息挖掘的研究较少 ， 主要是利用深度学习技术对图片内容进行识别 ， 进而分析城市意象要素类型城市街区的绿化指数 ［ １ Ｍ ２ １ 、 街道安全指数 ［ １ ３ ］ 等 ， 与城市物质形态 、 建成环境息息相关 。 另一方面 ， 就建成环境审计研究而言 ， 相较于传统的审计方法 ， 基于街景图片的虚拟审计以更广泛的地理覆盖范围 、 更高的性价比 、 更高的更新频率成为广泛运用的 、 有效且可靠的测度手段之一 。 而其主要的研究内容多集中在公共健康领域 ， 探究城市空间的要素和其品质对居民健康和活动的影响 ， 如邻里环境的活动友好性 １ １ ４ ＊ １ ５１ 、 促进不健康饮食的建成环境特征社区环境的致肥性 ｎ ７ １ 和街道清洁程度对城市健康的影响 ［ １ ８ ］ 等 ； 城市规划与设计领域则有对城市用地类型分类 ［ １ ９ ］ 、 街道可步行性 ［ ２ｌ ） １ 和城市空间品质等的探讨 ， 但少有直接对空间失序的现象进行研究与测度 ， 原因可能是空间失序的概念较新颖 、 抽象而评估方法较为模糊等 。 因而本研究希望探讨如下问题 ： （ １ ） 在中国城市空间特征的语境下 ， “ 空间失序 ” 准确的概念及其要素是什么？ （ ２ ） 空间失序在北京的城市空间中是否普遍存在？ （ ３ ） 如果存在 ， 相较于西方城市 ， 北京城市空间失序有怎样特殊的空间表征 ， 以及造成空间失序的原因和可能的影响是什么？ ２ 研究方法研究分为 ４ 个步骤进行 （ 见图 １ ） 。 首先基于对结粜分析评价 １ ． 空间失序的大规模测度研究步骤图示 １ ． Ｔｅｃ ｈ ｎ ｉ ｃ ａ ｌ ｒ ｏ ｕｔ ｅ 中国城市空间的整体现状认知 ， 开展了局部现场调研 ， 以进行信息搜集 ； 其次总结国外的城市案例与已有理论研究 ， 整合中国城市空间发展特征 ， 建立起空间失序的量化指标体系 ； 随后 ， 通过腾讯地图 Ａ Ｐ Ｉ 提取街道和路网数据 ， 由此获取海量的街景数据 ， 并依据指标体系建立在线审计平台 ， 由审计员对街景图片进行分要素的人工判读 ； 最后 ， 进行空间分析与计量分析 ， 得到街道的 “ 空间失序 ” 指数 ， 并探究城市空间失序对城市活力可能造成的影响 。 ２ ． １ 中国城市空间失序的构成要素识别本研究借鉴了已有西方城市研究中传统的建成环境审计方法 ， 即审计清单 （ ｃ ｈ ｅ ｃ ｋ ｌ ｉ ｓ ｔ ） 。

随着城市建设提质优化以及人们对美好生活的追求，空间品质成为城市研究中的重要组成部分。但近年来经济的高速发展，城市空间出现失序。以合肥市二环内区域为案例，以街景图像等多源数据为载体，采用非现场建成环境审计等技术方法， 探索合肥市空间失序现象以及不同类型的街道与空间失序程度的关系。 结果表明合肥市二环内： （1） 整体空间破败程度为 35.11%；（2）空间失序要素中以沿街商业要素的失序程度最为严重；（3）商业服务业设施用地街道（B 类）空间品质最差，物流仓储用地（W 类）街道空间品质优质。基于空间失序理论，大规模测度街道空间品质的优劣，在实践上能为未来城市的精细化管理提供重要依据；在理论研究上尝试弥补以往国内城市地理对空间失序研究的空缺。