文章基本信息

标题：Smart Parking Management System Using AI
本地全文：下载
作者：In Hwan Jung ; Jae Moon Lee ; Kitae Hwang 等
期刊名称：Webology
印刷版ISSN：1735-188X
出版年度：2022
卷号：19
期号：1
页码：4629-4638
DOI：10.14704/WEB/V19I1/WEB19307
语种：English
出版社：University of Tehran
摘要：This paper is aimed to introduce a smart parking lot management system using multiple cameras and artificial intelligence technique. When a vehicle enters a parking lot, it recognizes the vehicle number using embedded camera, tracks which parking space the vehicle is parked in, and updates parking space information. In addition, using a surveillance camera images, it has been also implemented to detect collision accidents that may occur while the vehicle is moving in the parking lot. Vehicle number recognition system uses OCR technique and is implemented on a Raspberry system. By managing the vehicle number recognized at the entrance of the parking lot as an Object ID, it was possible to effectively track the vehicle as a moving object inside the parking lot and finally identify the parking location. In order for accident detection, YOLO with CNN deep learning process is used. More than 500 possible collision images are trained in advance. Experimental results show that the detection accuracy of parking and accident detection increases as the number of training images increases. The accident detection needed more training images because it has more diversity. By using the smart parking system implemented in this paper, it is possible to effectively manage the vehicle's parking location, free space information and possible accidents. Using a cloud system, implemented system can provide drivers an integrated parking lot information over large areas.
关键词：Smart Parking;YOLO;Deep learning;Vehicle Tracking;Accident Detection