摘要
在项目实践基础上,使用机器视觉技术,对采集的图像采用中值滤波窗口滤除采集过程中的脉冲及椒盐噪声;使用自适应大津法分割图像,在对单元板在线检测时,由于工作台振动造成模组摆放位置与标准模板位置产生偏差,可采用Fomier-Mellin法进行图像配准,并对配准结果进行双线性插值法处理,结果表明,该方法能精确确定图像分割时的阈值,配准结果满足工厂检测要求,适用于工业现场快速实时测量系统,具有较高的推广价值。
Based on the practice in project,this article is using machine vision technology,using median filter window filter pulse and the pepper noise of image in acquisition process,using adaptive OTSU segmentation image.With line detecting cell plate,because of the table vibration,the cell board placement position may be have bias with the standard template position.Fomier-Mellin method can be used to image registration,and the registration results can be treated by bilinear interpolation.The results show： this method can accurately determine the threshold in image segmentation;the registration results can meet the requirements in the factory test;this method can be used for fast real-time measurement system in Industrial site;have high value of the promotion.
出处
《电子设计工程》
2011年第11期148-151,共4页
Electronic Design Engineering
作者简介
庞伟伟（1987-），男，安徽合肥人，硕士研究生。研究方向：在线检测及仪器智能化。
