摘要
先进制造系统中,机器视觉用于工件位姿的检测。结合先进的图像采集系统,以VC++.NET为开发平台,实现了对工件位姿的自动识别,同时提出了一种基于机器视觉的工件轮廓图像采集与识别的方法。主要分析了识别过程中灰度变换、二值图像和边缘检测几个难点,对检测结果进行了分析和比较。结果表明灰度变换和二值化能很大地提高对灰度图像的识别精度,在边缘检测过程中,Roberts算子检测水平和垂直的效果好于斜向边缘,定位精度高,但对噪声敏感,Sobel边缘检测算子则对灰度渐变和噪声较多的图像处理得较好。
Machine vision is used to detect the position and form of a work piece in advanced manufacturwere analyzed and compared. The result indicates that the gray level varying and two valued image can greatly raise the identification precision of the gray image. In the course of detecting the edge, the effect of detecting the horizontal and vertical edge is better than that of detecting the sloped one by Roberts algorithm, and the location precision is high, but it is easily affected by the noise, also the result of processing the image with gray gradations and much noise by Sobel algorithm is good.
出处
《电子工程师》
2006年第4期29-31,65,共4页
Electronic Engineer
关键词
机器视觉
灰度变换
二值图像
边缘检测
machine vision
gray level varying
two valued image
edge detection
