| 周雪,张舜尧,赵骏生.基于机器视觉与Andriod的停车场智能管理系统[J].井冈山大学自然版,2018,(6):39-43 |
| 基于机器视觉与Andriod的停车场智能管理系统 |
| INTELLIGENT PARKING MANAGEMENT SYSTEM BASED ON MACHINE VISION AND ANDRIOD |
| 投稿时间：2018-10-03 修订日期：2018-10-24 |
| DOI：10.3969/j.issn.1674-8085.2018.06.007 |
| 中文关键词: 停车场智能管理 机器视觉 Andriod系统 大数据样本 卷积神经网络 车牌检测 |
| 英文关键词: machine vision Andriod big data sample convolution neural network license plate detection |
| 基金项目:福建省中青年教师科研资助项目（JAT171206 |
|
| 摘要点击次数: 1921 |
| 全文下载次数: 2717 |
| 中文摘要: |
| 为了实现停车场的智能与精确管理，设计了一套基于机器视觉与Andriod的停车场智能管理系统。首先，集成网络摄像头、数据处理服务器硬件和Andriod终端显示器，组成系统硬件平台；嵌入机器视觉算法和局域网通信，构建了停车场智能管理系统架构。然后，基于10万帧各场景各车型的大数据样本，利用卷积神经网络四层训练，设计了车牌定位检测算子，实现车辆有无判断与车牌定位。在Visiual Studio平台开发系统，并对该智能管理系统进行了测试，输出结果表明：所提出的停车场智能管理系统，在车牌检测和系统智能性方面，都优于传统停车场管理系统。 |
| 英文摘要: |
| In order to achieve intelligent and accurate parking management, a parking intelligent management system based on machine vision and Andriodwas proposed in this paper.Firstly, the system hardware platform was integrated by network cameras, data processing server hardware and an Android terminal display. And the car park intelligent management system structure was constructed by embedding machine vision algorithm and LAN communication. Then, a license plate locating and detecting operator was designed by using convolution neural network four-layer training based on 100,000 frames of large data samples of each vehicle type in each scene to realize the vehicle identification and license plate location. The Visual Studio platform development systemwas used to test the proposed intelligent management system. The results show that the proposed intelligent parking management system is superior to the traditional parking management system in license plate detection and system intelligence. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
