二维机器视觉测量中的图元识别  

Image Element Recognition in 2D Machine Vision Measurement

作　　者：朱宏德[1] 娄小平[1] 邓文怡[1] 吕乃光[1] 

机构地区：[1]北京机械工业学院,北京100085

出　　处：《计算机工程与应用》2006年第3期217-218,共2页Computer Engineering and Applications

基　　金：北京市教委科技发展计划面上项目(编号:KM200311232133)

摘　　要：采用视觉测量方法对工件的各种断面尺寸参数进行测量时,需要在采集到的图像中找到被测量对象。文章给出了一种识别测量对象的图像匹配方法。该方法首先采用Canny方法从工件图像中提取出宽度为单像素的连续边缘图像,并根据工件的原理图纸生成待匹配的模板图像;对模板图像进行缩放和旋转,使之能在匹配率的约束下和边缘图像实现匹配;原理模板中包含图元的编码信息,据此可以在边缘图像中识别出工件的图元信息。实验结果表明,文中提出的方法能适用于一般工件的断面图元识别。It is necessary to get the measure elements in the grabbed image when measuring various dimension parameters of the workpiece＇s cross-section by machine-vision method.A matching method is presented to recognize the measure elements.The method gets the one-pixel-width edge image by Canny computational method and then creates a template image to be matched according to the workpiece＇s dimension description.Zooming and rotating the template image to match the edge image by the adjustment of match rate.The template image includes coding information of image elements by which certain image element can be recognized in the edge image.Resuh shows that the presented method can be used in many cross-section measurements.

关 键 词：图元识别 断面测量 二维机器 视觉测量 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]