机器视觉系统在拉链缺陷检测中的应用

摘要

受工艺过程影响，拉链在生产过程中会产生缺齿次品，传统方法是靠人眼和测量工具进行检测，耗费大量人力和时间。采取机器视觉技术，对拉链图像的特征进行研究，提出针对拉链齿数检测的三种检测原理：神经网络识别方法、模板匹配识别方法和基于一维灰度投影的识别方法。根据检测原理分别编写核心处理程序及其上位机软件，模拟工厂实际检测环境设计搭建实验平台，进行检测实验。%Defective goods with missing teeth are produced in the production process of zippers.TraditionaI detection approach depends on the human eyes and measuring tooIs which cost Iots of manpower and time.The machine vision technoIogy is taken to study the feature of zipper image.Three kinds of principIes are put forward:the neuraI network recognition method, the tempIate matching and the 1-dimensionaI gray statisticaI.Design the software and the experiment pIatform then experi-mentaIize the test on the pIatform.