基于计算机视觉的城市老旧小区环境质量大规模 测度研究* ——以武汉市为例
Large-scale Measurement of Environment Quality of Urban Old and Dilapidated Communities Based on Computer Vision: A Case Study of Wuhan
高非凡
中国人民大学公共管理学院 博士研究生
李志刚（通信作者）
武汉大学城市设计学院 湖北省人居环境工程技术研究中心 教授，博士生导师，zhigangli@whu.edu.cn
摘要: 在高质量发展时代，科学运用新数据、新技术开展城市老旧小区体检工作具有重大现实意义。以武汉市为例，基于安居 客小区实景图像，运用深度学习、空间计量等方法，从“市—区—街道”多尺度进行老旧小区体检评估，刻画老旧小区 环境质量的空间分布特征，揭示小区环境质量与房价的空间相关关系。研究发现：（1）武汉市小区环境质量呈现“主 城区由沿江两岸向外围地区逐渐升高，远城区高值围绕低值”的空间分布特征，老旧小区环境质量显著低于新建小区。 （2）低质量的老旧小区在沿江地区连片集聚，以江岸区南部和江汉区东南部的老里分小区、青山区的单位房小区、武昌 区的商品房和单位房小区为典型。（3）小区环境质量和房价存在显著的空间相关性，二者从中心向外围形成了“低— 高、高—高、高—低、低—低”的空间集聚格局。（4）小区环境质量是影响房价的关键因素，特别是在新建小区。由此为 城市小区环境质量的大规模自动化评估提供技术工具，对于快速识别亟待改造的重点区域具有重要启发。
Abstract: In the era of high-quality development, it is of great practical significance to scientifically apply new data and new technology to carry out physical examinations in urban old and dilapidated communities. Taking Wuhan as an example, based on Anjuke housing images, this paper applies deep learning and spatial econometric analysis to community evaluation from "city-district-street" scales, describing the spatial distribution characteristics of environment quality of urban old and dilapidated communities, and revealing the spatial correlation between community environment quality and housing price. The results show that: (1) Community environment quality in Wuhan gradually increases from the banks along the Yangtze River to the peripheral areas in urban districts, and high-value points are distributed around low-value points in suburban districts. Environment quality of old and dilapidated communities is significantly lower than that of newly-built communities. (2) Low-quality old and dilapidated communities gather along the Yangtze River. Typical examples include Lifen communities in the south of Jiang'an District and southeast of Jianghan District, unit communities in Qingshan District, and commercial housing and unit communities in Wuchang District. (3) There is a significant spatial correlation between community environment quality and housing price. The two form a spatial agglomeration pattern of "low-high, high-high, high-low, low-low" from the center to the periphery. (4) Community environment quality is the key factor that affects housing prices, especially in newly built communities. This study provides a technical tool for large-scale automatic evaluation of urban community environment quality. It also has important enlightenment for quickly identifying the key areas that need to be reformed.
关键词：老旧小区；小区环境质量；城市体检；图像回归；深度学习；空间计量；武汉市
Keyword: old and dilapidated community; community environment quality; city health examination; image regression; deep learning; spatial econometrics; Wuhan
中图分类号：TU984
文献标识码: A
吴志强，伍江，张佳丽，等. “城镇老旧小区更新
改造的实施机制”学术笔谈[J]. 城市规划学刊，
2021（3）：1-10.
WU Zhiqiang, WU Jiang, ZHANG Jiali, et al.
Symposium on implementation mechanism of
renewing old urban quarters[J]. Urban Planning
Forum, 2021(3): 1-10.
刘佳燕，张英杰，冉奥博. 北京老旧小区更新改
造研究：基于特征—困境—政策分析框架[J]. 社
会治理，2020（2）：64-73.
LIU Jiayan, ZHANG Yingjie, RAN Aobo. A study
on old community regeneration in Beijing: based
on characteristic-predicament-policy analytical
framework[J]. Social Governance Review, 2020(2):
64-73.
王越，陈烨. 从北京老旧小区物业失管看社区基
层组织困境[J]. 城市发展研究，2020，27（10）：
134-140.
WANG Yue, CHEN Ye. A study on the plight
of community grassroots organizations from
the perspective of property management loss in
Beijing old residential area[J]. Urban Development
Studies, 2020, 27(10): 134-140.
邬樱，李爱群. 北京老旧小区更新改造政策梳理
与柔性化策略研究——双重老龄化视角[J]. 城
市发展研究，2022，29（5）：73-79.
WU Ying, LI Aiqun. Research on renovation and
reconstruction policies and flexibility strategies
of old residential areas in Beijing: a dual aging
perspective[J]. Urban Development Studies, 2022,
29(5): 73-79.
赵慧敏，龙瀛. 中国居住区外观空间质量指标
体系构建及其平顶山应用研究[J]. 城市设计，
2022（2）：48-55.
ZHAO Huimin, LONG Ying. Research on the
construction of index system for residential
appearance space quality in China and its application in Pingdingshan[J]. Urban Design, 2022(2):
48-55.
伍江，王信，陈烨，等. 超大城市城市体检的挑战
与上海实践[J]. 城市规划学刊，2022（4）：28-34.
WU Jiang, WANG Xin, CHEN Ye, et al. The
difficulties of megacities' city health examination
and the practice in Shanghai[J]. Urban Planning
Forum, 2022(4): 28-34.
王宇凡，柴康妮，卓云霞，等. 中国城市流动人口
住房质量的空间分异与影响因素[J]. 地理学报，
2021，76（12）：2944-2963.
WANG Yufan, CHAI Kangni, ZHUO Yunxia,
et al. Spatial variation of migrant population's
housing quality and its determinants in China's
prefecture-level cities[J]. Acta Geographica Sinica,
2021, 76(12): 2944-2963.
林李月，朱宇，梁鹏飞，等. 基于六普数据的中国
流动人口住房状况的空间格局[J]. 地理研究，
2014，33（5）：887-898.
LIN Liyue, ZHU Yu, LIANG Pengfei, et al. The
spatial patterns of housing conditions of the floating
population in China based on the sixth census
data[J]. Geographical Research, 2014, 33(5): 887-
898.
KEALL M, BAKER M G, HOWDEN-CHAPMAN
P, et al. Assessing housing quality and its impact
on health, safety and sustainability[J]. Journal of
Epidemiology and Community Health, 2010, 64(9):
765-771.
SUGLIA S F, DUARTE C S, SANDEL M T.
Housing quality, housing instability, and maternal
mental health[J]. Journal of Urban Health, 2011,
88(6): 1105-1116.
GIELEN A C, WILSON M E H, FADEN R R, et
al. In-home injury prevention practices for infants
and toddlers: the role of parental beliefs, barriers,
and housing quality[J]. Health Education &
Behavior, 1995, 22(1): 85-95.
杨静，吕飞，史艳杰，等. 社区体检评估指标体系
的构建与实践[J]. 规划师，2022，38（3）：35-44.
YANG Jing, LYU Fei, SHI Yanjie, et al. Establishing
community physical examination evaluation index
system and its practice[J]. Planners, 2022, 38(3):
35-44.
张乐敏，张若曦，黄宇轩，等. 面向完整社区的城
市体检评估指标体系构建与实践[J]. 规划师，
2022，38（3）：45-52.
ZHANG Lemin, ZHANG Ruoxi, HUANG
Yuxuan, et al. Urban physical examination evaluation index system research and practice towards
integrated community development[J]. Planners,
2022, 38(3): 45-52.
何宛余，李春，聂广洋，等. 深度学习在城市感知
的应用可能——基于卷积神经网络的图像判别
分析[J]. 国际城市规划，2019，34（1）：8-17.
HE Wanyu, LI Chun, NIE Guangyang, et al.
The latent application of deep learning in urban
perception: image discrimination analysis by
convolutional neural network[J]. Urban Planning
International, 2019, 34(1): 8-17.
何炬，张文忠，曹靖，等. 多源数据在城市体检中
的有机融合与应用——以北京市为例[J]. 地理
科学，2022，42（2）：185-197.
HE Ju, ZHANG Wenzhong, CAO Jing, et al.
Organic integration and application of multi-source
data in city health examination: a case of Beijing[J].
Scientia Geographica Sinica, 2022, 42(2): 185-197.
刘智谦，吕建军，姚尧，等. 基于街景图像的可解
释性城市感知模型研究方法[J]. 地球信息科学
学报，2022，24（10）：2045-2057.
LIU Zhiqian, LYU Jianjun, YAO Yao, et al.
Research method of interpretable urban perception
model based on street view imagery[J]. Journal of
Geo-information Science, 2022, 24(10): 2045-2057.
HELBICH M, YAO Y, LIU Y, et al. Using deep
learning to examine street view green and blue spaces
and their associations with geriatric depression in
Beijing, China[J]. Environment International, 2019,
126: 107-117.
YAO Y, XU C, YIN H, et al. More visible greenspace, stronger heart? Evidence from ischaemic
heart disease emergency department visits by
middle-aged and older adults in Hubei, China[J].
Landscape and Urban Planning, 2022, 224: 104444.
吴佳雨，王诗奕，李红，等. 基于人工智能技术的
植被色彩感知与焦虑关联分析[J]. 地理学报，
2023，78（4）：1044-1056.
WU Jiayu, WANG Shiyi, LI Hong, et al. The
correlation between plant color perception
and anxiety based on artificial intelligence
technology[J]. Acta Geographica Sinica, 2023,
78(4): 1044-1056.
胡一可，张天霖，王磊，等. 景观服务视角下城市
街区感知测度及空间分布特征[J]. 风景园林，
2022，29（10）：45-52.
HU Yike, ZHANG Tianlin, WANG Lei, et al.
Measurement and spatial distribution of perception
in urban blocks from the perspective of landscape
service[J]. Landscape Architecture, 2022, 29(10):
45-52.
陈婧佳，张昭希，龙瀛. 促进公共健康为导向的
街道空间品质提升策略——来自空间失序的视
角[J]. 城市规划，2020，44（9）：35-47.
CHEN Jingjia, ZHANG Zhaoxi, LONG Ying.
Strategies for improving the quality of urban
street space oriented to promoting public health:
perspective from spatial quality[J]. City Planning
Review, 2020, 44(9): 35-47.
龙瀛，唐婧娴. 城市街道空间品质大规模量化
测度研究进展[J]. 城市规划，2019，43（6）：
107-114.
LONG Ying, TANG Jingxian. Large-scale
quantitative measurement of the quality of urban
street space: the research progress[J]. City Planning
Review, 2019, 43(6): 107-114.
狄迪，蒋应红，叶丹，等. 新城市科学支持下的城
市街道空间品质综合评价——以上海市中心城
区为例[J]. 规划师，2021，37（16）：5-12.
DI Di, JIANG Yinghong, YE Dan, et al. Street
space quality evaluation supported by new urban
science: the central district of Shanghai[J].
Planners, 2021, 37(16): 5-12.
LI S, MA S, TONG D, et al. Associations between
the quality of street space and the attributes of the
built environment using large volumes of street
view pictures[J]. Environment and Planning B:
Urban Analytics and City Science, 2022, 49(4):
1197-1211.
张书杰，李文竹，龙瀛，等. 基于多年街景图片的
城市街道步行设施改善评价——以中国45个城
市为例[J]. 城市发展研究，2022，29（6）：53-64.
ZHANG Shujie, LI Wenzhu, LONG Ying, et al.
Evaluation on the improvement of pedestrian
facilities with multi-year street view images[J].
Urban Development Studies, 2022, 29(6): 53-64.
刘颖，袁媛，邢汉发，等. 街景图片识别城市贫困
的适用性——基于广州市中心城区的验证[J].
热带地理，2020，40（5）：919-929.
LIU Ying, YUAN Yuan, XING Hanfa, et al. The
applicability of street view images to identify
urban poverty in the central urban region of
Guangzhou[J]. Tropical Geography, 2020, 40(5):
919-929.
向雨，张鸿辉，刘小平. 多源数据融合的城市体
检评估——以长沙市为例[J]. 热带地理，2021，
41（2）：277-289.
XIANG Yu, ZHANG Honghui, LIU Xiaoping.
Urban physical examination evaluation based on
multisource data: a case study of Changsha City[J].
Tropical Geography, 2021, 41(2): 277-289.
郭金函，马子迎，边经卫，等. 基于街景图像的厦
门本岛环境特征对住宅价格的影响研究[J]. 地
球信息科学学报，2022，24（11）：2128-2140.
GUO Jinhan, MA Ziying, BIAN Jingwei, et
al. Analysis on the influence of environmental
characteristics of Xiamen island on housing price
based on street view imagery[J]. Journal of Geoinformation Science, 2022, 24(11): 2128-2140.
XU W, GU Y, CHEN Y, et al. Combining deep
learning and crowd-sourcing images to predict
housing quality in rural China[J]. Scientific
Reports, 2022, 12: 19558.
龙瀛. 颠覆性技术驱动下的未来人居——来自
新城市科学和未来城市等视角[J]. 建筑学报，
2020（s1）：34-40.
LONG Ying. Future of human habitats driven by
disruptive technologies: perspectives from the new
science of cities and future cities[J]. Architectural
Journal, 2020(s1): 34-40.
FLEISS J, COHEN J. The equivalence of weighted
kappa and the intraclass correlation coefficient
as measures of reliability[J]. Educational and
Psychological Measurement, 1973, 33(3): 613-619.
付国栋，黄进，杨涛，等. 改进CBAM的轻量级注
意力模型[J]. 计算机工程与应用，2021，57（20）：
150-156.
FU Guodong, HUANG Jin, YANG Tao, et al.
Improved lightweight attention model based
on CBAM[J]. Computer Engineering and
Applications, 2021, 57(20): 150-156.
ANSELIN L. Local indicators of spatial
association-LISA[J]. Geographical Analysis, 1995,
27(2): 93-115.
李博闻，黄正东，刘稳. 基于公交服务需求与供
给匹配程度的公交站点布局评价——以武汉市
为例[J]. 现代城市研究，2019（5）：99-105.
LI Bowen, HUANG Zhengdong, LIU Wen.
Assessment of the bus stop locations based on the
coupling level between demand and supply of bus
service: a case study of Wuhan[J]. Modern Urban
Research, 2019(5): 99-105.
姜磊. 空间回归模型选择的反思[J]. 统计与信息
论坛，2016，31（10）：10-16.
JIANG Lei. The choice of spatial econometric
models reconsidered in empirical studies[J].
Journal of Statistics and Information, 2016,
31(10): 10-16.
宋伟轩，马雨竹，陈艳如. 南京城区住宅售租
价格时空分异与影响因素[J]. 地理科学进展，
2018，37（9）：1268-1276.
SONG Weixuan, MA Yuzhu, CHEN Yanru.
Spatiotemporal differentiation and influencing
factors of housing selling and rental prices: a case
study of Nanjing City[J]. Progress in Geography,
2018, 37(9): 1268-1276.
