摘要
为解决教室照明用电浪费问题,设计和制作了基于机器视觉技术的教室照明节能控制系统。该系统在监控计算机端采用帧差分与分段收敛的背景差分叠加法从监控录像中识别教室人员,并根据室内人员情况控制教室照明设备的开启或关闭。系统硬件主要由图像采集模块、监控计算机、电气控制模块及教室电气线路等4部分组成;系统软件基于Matlab开发,按功能可分为图像获取、图像处理、串口双向传输、图形用户界面及数据库等5部分。室内人员检测实验结果表明,系统识别平均耗时0.2s,进行实时室内人员识别的帧速上限为5帧/s;当图像中不包含和包含人员时,系统对室内人员的检测准确率分别为100%和93%。
In order to solve the energy waste problem of classroom lightening, a classroom lightening energy-saving control sys tem based on machine vision techniques is designed and developed. People inside a classroom could be identified from surveillance video using the superposition of frame difference and sectional convergence based background difference algorithms at the system＇s monitoring computer. Lightening equipments within the classroom could be lightened or shut down according to the room personnel situation. The system hardware is composed of image acquisition module, monitoring computer, electrical control module and classroom electrical wiring. The system software is developed using MATLAB and could be divided into five parts as image acquisition, image processing, serial transmission, graphical user interface and data base according to their functions. Room personnel detection accuracy experimental results indicate that： averaged system detection time is 0.2s, the frame speed limit for personnel detection is five frames per second, when the video frame did not contain and included personnel, the system＇s room personnel detection accuracy is 100~ and 93~, respectively.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第4期1648-1652,共5页
Computer Engineering and Design
基金
广东省科技计划基金项目(2011B020308009)
华南农业大学博士启动基金项目(4500-K10056)
关键词
机器视觉
节能控制
照明控制
图像处理
视频监控
machine vision
energy-saving control
lightening control
image processing
video surveillance
