汽车指针仪表的视觉检测系统的研究与设计  

Research and Design of Vision Inspection System of Automobile Pointer Instrument

作　　者：徐洋[1] 廖钦渔[1] 李锐[1] 程安宇[1] 

机构地区：[1]重庆邮电大学汽车电子与嵌入式系统工程研究中心,重庆400065

出　　处：《电视技术》2012年第17期139-143,共5页Video Engineering

基　　金：重庆高校创新团队建设计划项目(201014);重庆市科技攻关计划项目(CSTC2008AA6025);重庆高校优秀成果转化资助项目(KJZH11207);重庆矢崎仪表有限公司"汽车智能网络仪表测试系统平台研发"项目(渝发改技函[2011]424号)

摘　　要：研究了一种基于机器视觉的汽车仪表盘自动检测系统。该系统通过机器视觉的方法,采用OpenCV编程环境,采集有效的汽车指针仪表盘图像,并对其进行处理,识别出仪表刻度值与指针的位置和重合程度,从而判定仪表指示准确性是否符合要求。系统包含了光源的配置、摄像头的选取和其横纵竖三向的移动、图像的预处理以及图像的识别判定等部分。测试证明,该系统能有效地自动读出指针仪表的示数并判断其是否准确,识别速度和精确度与国外类似产品相比得到了提升。A system of automobile instrument inspection is designed, which is based on the OpenCV programming environment and uses machine vision technology. At first, the system collects effective images of the automobile instrument, and then the system will do some necessary image processing and recognition in order to readout the value of the instrument calibrations. Through those the system could judge the accuracy of the instrument whether it meets the requirements or not. The design contains several parts such as configuration of light source, selection and horizontal longitudinal and vertical movement of camera, image preprocessing and image recognition and so on. Experiments show that, this system can effectively read out the value of the vehicle pointer instrument and points out the dial precision by an automatic way, comparing with the similar products abroad the distinguishing speed and accuracy have obvious improvement.

关 键 词：机器视觉 指针仪表 OPENCV 

分 类 号：TN911.93[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]