摘要
研究了一种基于机器视觉的带钢表面缺陷检测系统 ,它采用模块化硬件设计 ,图像处理软件满足实时检测的要求 ,可以有效地检测出生产线上的带钢表面缺陷 .为该系统设计了一种基于规则表分类器、模糊算法及人工神经网络的组合式多级分类器 ,具有一定的学习能力 ,当待测材料或有关设备发生变化时 ,系统可以根据缺陷样本库对分类器进行训练 ,以适应生产线的相关变化 .系统具有较强的容错性。
A steel strip surface inspection system based on computer vision was presented which uses modularized frame of hardware. The software of image processing can detect and classify defects across the entire steel strip surface. The system develops an effective assembled classifier by several pattern recognition technologies including rule based table, fuzzy algorithm and neural network. It has the ability of self learning and can adapt to different equipments and materials by the training of classifier according to different sets of samples. The performance and robustness of the system is demonstrated by the experiments in steel strip production line in Bao Steel Company.
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2003年第2期72-74,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家经贸技术创新计划项目