首页|基于全景图像的社区体检评估技术应用研究

基于全景图像的社区体检评估技术应用研究

扫码查看
2023年住建部全国城市体检工作将城市体检单元细化到了住房、小区(社区)、街区、城区(城市)四个尺度,其中住房、小区(社区)、街区尺度的体检工作需要通过实地调研采集来开展,且体检要求更加精细化.为了提升调研采集效率和数据质量,减低体检评估成本,笔者所在的城市象限团队提出了基于全景图像的社区评估技术方案,通过智能感知背包和电动轮椅的一体化采集装备,完成体检调研过程中全景图像的自动采集;并结合图像算法、物联数据处理和空间分析技术进行云端数据智能处理,初步实现了包括环境品质、路面平整、无障碍设施、公共空间品质等50多项建成环境指标的智能分析计算,探索了基于智能感知的全景图像进行体检指标计算的评估方法.
Application Research on the Assessment Technology of Community Physical Examination Based on Panoramic Images
The 2023 National Urban Physical Examination of the Ministry of Housing and Construction has re-fined the urban physical examination unit to four scales:housing,district(community),neighbourhood,and urban area(city),of which the physical examination of housing,district(community),and neighbourhood scales needs to be carried out through on-site research and collection,and the physical examination require-ments are more refined.In order to improve the research collection efficiency and data quality,and reduce the cost of physical examination and assessment,the author's Urban Quadrant team proposed a community as-sessment technology programme based on panoramic images,through the integration of intelligent perception backpacks and electric wheelchairs collection equipment,to complete the automatic collection of panoramic images during the physical examination and research process;and combined with image algorithms,Internet of Things(IoT)data processing and spatial analysis technology for cloud-based data intelligent processing,initial The intelligent analysis and calculation of more than 50 built environment indicators,including environ-mental quality,road surface levelling,barrier-free facilities,public space quality,etc.,have been achieved,and the assessment method of calculating medical check-up indicators based on the panoramic image of intelligent perception has been explored.

City checkupIntelligent perceptionComputer vision

张孝贤、周元婧、李大勇、姜冬睿、茅明睿

展开 >

北京城市象限科技有限公司

北京社区研究中心

城市体检 智能感知 计算机视觉

2024

建筑师
中国建筑工业出版社

建筑师

影响因子:0.374
ISSN:1001-6740
年,卷(期):2024.(2)
  • 30