多指针式指示仪表识别模块的设计与实现

发布时间：2018-02-14 04:14

本文关键词： 机器视觉多指针式仪表图像处理技术模式匹配识别模块　出处：《内蒙古大学》2017年硕士论文　论文类型：学位论文

【摘要】：百分表由于其使用方便,抗干扰能力强,目前已经广泛应用于实际生产和生活中。它的功能主要是用来测量微小位移量,长时间的高频率使用,会使百分表产生误差。产品的质量和安全与百分表的准确度有着密不可分的关系,因此对百分表的精度检定是非常重要的环节。根据国家计量部门制定的专门检定规程,百分表的检定周期一般不超过1年。现如今百分表的检定工作大多数都是检定人员长时间用肉眼观察仪表的读数并记录数据,这种高强度长时间地观察容易使人眼感到疲惫,并且引入不确定的人为误差,检定精度受到影响。本文设计的多指针式指示仪表识别模块包括图像的采集系统,图像预处理系统以及仪表读数识别系统。该识别模块运用工业相机、光学镜头、照明光源等设备搭建采集系统硬件平台,并利用计算机软件设计实现表盘图像处理和读数识别。机器视觉代替人眼技术与图像自动处理技术相结合,将图像信息转换成数字信息从而得到仪表的读数。将该识别模块应用在百分表全自动检定系统中,消除人眼读数视觉疲劳的同时还可以提高读数效率和精度。多指针式指示仪表识别模块中仪表图像应用CMOS型工业相机采集,通过USB传输线将仪表图像传输给计算机。图像预处理系统做进一步预处理操作。论文中仪表的读数识别系统中采用了两种算法。霍夫变换识别算法是在MATLAB软件平台验证实现,该算法在预处理仪表图像上检测指针直线,通过计算得出直线的斜率,利用斜率求出直线夹角,从而得出仪表示值。模式匹配识别算法是在LabVIEW软件平台验证实现,这种算法先将仪表盘上感兴趣区域提取出来作为模板,通过LabVIEW中模式匹配子模块进行像素匹配。通过确定表盘零刻度位置、仪表指针位置和表盘圆心位置得出指针旋转过的角度,最终得出读数。通过实际验证,模式匹配识别算法满足设计要求,最大识别误差不超过正负0.003mm,可应到百分表检定系统中。本系统还设计了友好的人机界面,与百分表检定系统完成衔接配合,为使用者提供方便。
[Abstract]:Because of its convenient use and strong anti-interference ability, the percentile meter has been widely used in practical production and daily life. Its function is mainly used to measure the small displacement and the use of high frequency for a long time. The quality and safety of the product are inextricably related to the accuracy of the percentile, so it is very important to verify the accuracy of the percentile. The verification period of a percentile is generally less than one year. Nowadays, most of the verification work of a percentile is made by the examiners who use the naked eye for a long time to observe the readings of the meter and record the data. This high intensity and long observation easily makes the human eye tired. The precision of verification is affected by the introduction of uncertain human error. The identification module of multi-index indicator instrument designed in this paper includes the image acquisition system. The image preprocessing system and the instrument reading recognition system are used to build the hardware platform of the acquisition system using industrial camera, optical lens, lighting source, etc. The computer software is used to realize the image processing and reading recognition of the dial. The machine vision technology is combined with the automatic image processing technology instead of the human eye. The image information is converted into digital information to get the reading of the instrument. The recognition module is applied in the automatic verification system of the percentile. At the same time, the reading efficiency and precision can be improved. The instrument image is collected by CMOS industrial camera in the identification module of multi-index indicator instrument. The instrument image is transmitted to the computer by USB transmission line. The image preprocessing system makes further preprocessing operation. In this paper, two algorithms are adopted in the instrument reading recognition system. The Hough transform recognition algorithm is verified and implemented on the MATLAB software platform. The algorithm detects the pointer line on the image of the preprocessing instrument, calculates the slope of the line, calculates the angle of the line by using the slope, and then obtains the indication value of the instrument. The pattern matching recognition algorithm is verified and implemented on the LabVIEW software platform. In this algorithm, the region of interest is extracted from the dashboard as a template, and the pixel matching is carried out through the pattern matching submodule in LabVIEW. The zero scale position of the dial is determined by determining the location of the zero scale of the dial. The pointer position of the meter and the center position of the dial are obtained by the rotation angle of the pointer, and the final reading is obtained. Through the actual verification, the pattern matching recognition algorithm meets the design requirements. The maximum recognition error is not more than plus or minus 0.003 mm, and it can be applied to the verification system of the percentile. The system also designs a friendly man-machine interface, which is connected with the verification system of the percentile meter and provides convenience to the user.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】