机器视觉仪表识别方法的研究进展  

Research Progress of Machine Vision Instrument Recognition

作　　者：黄颖怡[1] 汪仁煌[1] 岳利军[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510090

出　　处：《自动化仪表》2009年第8期58-60,共3页Process Automation Instrumentation

摘　　要：介绍了机器视觉仪表识别方法的研究进展,具体给出了仪表刻度线、指针的定位、读数识别的原理和方法,并将这些图像处理算法的优劣进行对比,提出要结合仪表特征(多指针多颜色、均匀刻度、非均匀刻度)选择算法进行识别;同时,分析了采集设备中摄像头、光源对后续图像处理过程造成的影响;最后总结仪表识别的研究状况及存在的问题,为实现实验室型高准确度仪表的机器视觉自动化校表给出了一些参考建议。The research progress of machine vision instrument recognition is introduced. The recognition principle and method tor instrument scaling, pointer location, and instrument reading are presented concretely, and the contrast of merits and demerits of various image processing algorithms is given. It is proposed thai the algorithm shall be selected based on instrument characteristics （multiple pointers in colors, uniform scales or non-uniform scales） for recognition; in addition, the influence of camera in collection equipment, light source on image processing is analyzed. The current research status and the problems existed are concluded; and suggestions are given for implementing highly precise machine vision automatic instrument calibration in laboratories.

关 键 词：指针式仪表 位置识别 读数识别 机器视觉 刻度 图像处理 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]