计算机视觉技术在微小尺寸测量中的应用  

Study on Measurement of Micro Size Based on Computer Vision Technology

作　　者：陈廉清[1] 袁红彬[2] 王龙山[2] 

机构地区：[1]宁波工程学院机械工程学院 [2]吉林大学机械科学与工程学院

出　　处：《农业机械学报》2006年第9期174-176,共3页Transactions of the Chinese Society for Agricultural Machinery

基　　金：宁波市工业科研攻关项目(项目编号:2005B100014)

摘　　要：应用计算机视觉技术快速测量微小零件的尺寸,研究了CCD图像预处理、图像二值化、边缘处理、标定等技术,发现中值滤波平滑处理图像的精度更高,轮廓跟踪能使图像边缘确定时的运算更简单。图像检测结果与工具显微镜测量进行了比较,结果表明该测量方法可以快速、准确地实现大批量同类零件的尺寸测量及合格度的检验,并可满足100%在线检测要求,检测误差不超过0.0119mm,每件检测时间不超过2s。In this paper the computer vision technology was applied to the rapid measurement. Based on the study of CCD images advance processing technology, the methods of binary image, edge processing as well as calibration were studied. It is pointed out that the precision of image in median filter smoothing algorithms is higher and the detection of image edge in contour following algorithms is more convenient. The paper also compared the result of image measurement with microscope measurement. The results showed that the former can complete the size measurement rapidly and accurately, and it also can get one hundred percent on-line detection. The measurement error will not exceed 0. 011 9 mm, and the measuring time is not more than 2 s for each part.

关 键 词：计算机视觉 尺寸测量 轮廓跟踪 中值滤波 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TG806[自动化与计算机技术—计算机科学与技术]