城市建筑2024，Vol.21Issue(11)：118-122,5.DOI:10.19892/j.cnki.csjz.2024.11.25
基于街景目标检测的城市人群活动研究
Research on Urban Crowd Activity Based on Street View Object Detection:A Case of Hanzheng Street in Wuhan City
摘要
Abstract
In recent years,the linkage between subjects such as machine learning and computer vision and other disciplines has become increasingly frequent,which also plays a springboard role in the study on architecture and urban space.In this paper,Mask-RCNN object detection algorithm is used to conduct image analysis and statistical research on the crowd of"cart workers"in the street view image of Hanzheng Street area in Wuhan City.The number of targets detected in the global street view is counted,and the activity characteristics of the crowd of"cart workers",the traffic pressure they face and the urban space usage problems reflected are studied.It is proposed that the data and visualization results of crowd activities can guide the selection of key areas in urban design.At the same time,the tag of"man-thing"should be added to the detection model used for urban research,so as to be suitable for complex population activity research.关键词
街景数据/Mask-RCNN目标检测/城市空间/人群活动/武汉"板车工"Key words
street view data/Mask-RCNN target detection/urban space/crowd activity/Wuhan "cart worker"分类
建筑与水利引用本文复制引用
吴子悦,孙靓,尚伟..基于街景目标检测的城市人群活动研究[J].城市建筑,2024,21(11):118-122,5.
