指针式仪表总体最小二乘图像校验算法  

Image Calibration Algorithm Based on Total Least Squares Method for Pointer Instruments

作　　者：张远辉[1] 张鼎[1] 许昌[1] 谢波[1] 

机构地区：[1]中国计量学院机电工程学院,浙江杭州310018

出　　处：《自动化仪表》2015年第5期75-79,共5页Process Automation Instrumentation

基　　金：国家自然科学基金资助项目(编号:61302191);浙江省教育厅科研基金资助项目(编号:Y201224542)

摘　　要：为了解决传统指针式仪表人工校验准确率低的问题,提出了一种基于计算机视觉的指针式仪表校验算法。将仪表图像由RGB颜色空间模型转换到HSV颜色空间模型,利用仪表盘指针的颜色特征信息提取指针;并采用总体最小二乘法拟合指针中心轴直线;通过腐蚀膨胀提取仪表刻度,计算指针与刻度的偏差,从而判断仪表是否合格。校验试验结果表明,该算法可靠性高、鲁棒性强、校验准确率高。To solve the problem of low accuracy for traditional pointer instruments caused by manual calibration, the verification algorithm based on computer vision for pointer instrument is proposed. The image of the instrument is transformed from RGB color space model into HSV color space model, the pointer is extracted by adopting the color characteristics of the panel pointer, then the total least squares method is used to fit the central axis of pointer; and the scale of the instrument is extracted through eroding and dilating algorithm, the deviation between the pointer and scale is calculated, thus to judge if the instrument is qualified or not. The result of calibration experiments indicates that this algorithm is reliable and robust; the calibration accuracy is high.

关 键 词：指针式仪表 计算机视觉 HSV颜色空间模型 总体最小二乘法 点到线距离 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]