机械零件二维几何尺寸和形状检测系统研究与开发  

Mechanical parts and shape of two-dimensional geometry detection system development

作　　者：严慧敏[1,2] 何炳蔚[2] 

机构地区：[1]福建信息职业技术学院机电工程系,福建福州350003 [2]福州大学机械工程及自动化学院,福建福州350002

出　　处：《机械》2008年第10期70-72,共3页Machinery

摘　　要：针对传统接触式尺寸测量方法的缺点,探讨利用单目视觉测量系统对被测工件进行尺寸测量,为产品的尺寸测量提供实时、快速、有效、经济的测量途径。首先,视觉系统获取零件图像,并对图像进行滤波处理;然后,提出采用Canny六阈值法进行边缘提取,得到完整的边缘图像;最终,利用最小二乘法拟合出机械零件的边缘形状,并通过计算获得零件的几何尺寸。实验表明该系统所得到的零件边缘图像清晰,多个几何尺寸的测量精度符合要求。与传统测量方法相比提高了检测零件的效率和精度,大大扩展了测量范围,从而充分体现了机器视觉测量的优点。Because of the drawbacks of the traditional touch measurement method, a new non-touch method with the use of Single CCD Vision system is presented, which provides a real time, fast, validity way. Firstly, the image of the parts is obtained by the vision system, and processed by using filter for clearing up the image noise. And then, six Thresholds method based on canny algorithm is presented so that the complete image of the object boundaries is acquired. At last, the edge of the mechanical parts is obtained by using least squares fitting, so, the geometries sizes of the parts are computed. Experiments show that the edge of image is clear and the measurement accuracy of the parts geometric dimensions meet the requirements in this system. Compared with traditional measurement methods, the detection efficiency and precision parts is improved and the range of measurement is much greater. So, the measurement advantages of machine vision are fully reflected.

关 键 词：单目视觉 几何尺寸 形状 边缘检测 机器视觉 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]