城市停车场车位信息视频检测方法  

Video Detection Method for Parking Space Information in Urban Parking Lots

作　　者：马寅晨 姜岚[1] 周煜申[1] 钱小聪 郭豆 武宜广[2] MA Yinchen;JIANG Lan;ZHOU Yushen;QIAN Xiaocong;GUO Dou;WU Yiguang(MCC Huatian Engineering&Technology Co.,Ltd.,Nanjing 210019,China;School of Geography,Nanjing Normal University,Nanjing 210023,China)

机构地区：[1]中冶华天工程技术有限公司,江苏南京210019 [2]南京师范大学地理科学学院,江苏南京210023

出　　处：《测绘与空间地理信息》2022年第7期36-38,43,共4页Geomatics & Spatial Information Technology

基　　金：安徽省重点研究与开发计划(202004a06020026)资助。

摘　　要：城市停车场车位信息的自动检测对于解决城市“停车难”问题和停车场智能化管理至关重要。但是目前的车位信息自动检测方法难以获取空置车位的地理位置信息。因此,本文提出一种停车场车位信息视频检测方法,该方法结合基于深度学习的车辆检测方法和视频空间化方法,利用停车场监控视频实现空置车位的数量信息和地理位置信息的自动检测。然后通过一个案例验证了本文方法的可行性。Automatic detection of parking space information in urban parking lots is key to solving the problem of"difficult parking"in cities and intelligent management of parking lots.However,the current automatic detection method of parking space information is difficult to obtain the geographical location information of vacant parking spaces.Therefore,a parking space information video detection method is proposed in this paper,which combines a deep learning-based vehicle detection method and a video spatialization method to achieve automatic detection of the quantity information and geographic location information of vacant parking spaces using parking lot surveillance video.Then the feasibility of the method in this paper is verified by a case study.

关 键 词：停车场 监控视频 车位信息 深度学习 视频空间化 

分 类 号：P208[天文地球—地图制图学与地理信息工程]