基于轮廓拟合与径向分割的指针式仪表自动读数方法

A recognition method of pointer instrument based on contour fitting and radial segmentation

针对指针式仪表人工读数劳动强度较大的问题,本文提出了一种基于轮廓拟合与径向分割的指针式仪表自动读数方法。该方法利用仪表的轮廓图像,结合常见的仪表盘刻度分布特性,提出轮廓分离法得到仪表盘刻度像素集,并采用最小二乘法拟合出表盘圆。在此基础上,通过分析指针的像素特性,提出径向指针分割算法,实现了仪表盘内指针区域的提取。根据指针偏转角度、零刻度和量程精度进行自动读数。为了验证算法的有效性,设计多组测试实验,结果表明本方法在读数误差允许范围内,能够克服宽窄指针、拍摄距离和表盘背景等因素的干扰,具有较强的自适应能力。

Aiming at the problem of labor intensity for manual reading of pointer instrument,this paper presents an automatic reading method of pointer instrument based on contour fitting and radial segmentation.The method is based on the instrument image which is contour extracted.Combined with the common distribution characteristics of the instrument panel,the pixel of instrument arc scale is obtained by the contour separation method,and the dial panel fitting circle is calculated by the least square method.On this basis,a radial segmentation algorithm is proposed to extract the pointer region in the instrument panel by analyzing the pixel characteristics of the pointer,and finally to realize the automatic reading based on pointer deflection angle,zero scale and rang accuracy.In order to verify the validity of the algorithm,multiple experiments are designed.The tset results show that the method is within the tolerance of reading error,and it can overcome impact of wide and narrow pointer,shooting distance,dial interference and so on,and has strong adaptive adjustment ability.