CAJ | 학술논문

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。
위료해결전통지침식의표인공교험준학솔저적문제，제출료일충기우계산궤시각적지침식의표교험산법。장의표도상유RGB안색공간모형전환도HSV안색공간모형，이용의표반지침적안색특정신식제취지침；병채용총체최소이승법의합지침중심축직선；통과부식팽창제취의표각도，계산지침여각도적편차，종이판단의표시부합격。교험시험결과표명，해산법가고성고、로봉성강、교험준학솔고。
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.