摘要
针对粉末冶金零件人工外观缺陷检测精度差、效率低、实时性不佳等问题,利用机器视觉实现自动化生产线粉末冶金零件缺陷的实时检测。系统选用工业相机采集零件图像,利用图像处理软件对零件图形进行特征提取,使用LIBSVM进行数据集训练,得到合格零件与缺陷零件的特征信息,实现零件的有效分类。这一系统检测精度高,可靠性好,能满足生产线实时性的要求。
Aiming at the problems of poor precision, low efficiency and poor real-time property of manual detection of appearance defects of powder metallurgy parts, the powder metallurgy parts in automatic production line can be detected by machine vision in real-time. The system selects the industrial camera to collect the part image, adopts the image processing software to carry on the feature extraction from the part image and uses the LIBSVM for the database training , the attributive information of the qualified parts and the defective parts can be obtained for effective classification of the parts. The system features high precision and reliability, and can meet the real-time requirement of the production line.
出处
《机械制造》
2017年第1期82-84,104,共4页
Machinery
关键词
机器视觉
粉末冶金零件
图像处理
缺陷检测
Machine Vision Powder Metallurgy Parts Image Processing Defect Detection
