笔管缺陷自动化检测系统设计与研究

中国已成为笔类产品生产大国与出口大国,而笔管检测是制笔行业的关键工艺技术。针对目前制笔行业中笔管检测的需求,设计了笔管缺陷自动化检测系统,以提高笔管缺陷检测效率及笔管制造企业的生产质量。基于机器视觉及重心分类装置,采取分模块检测系统,对笔管的缺陷形态、类型进行鉴别与统计,高效率、高精度地实现笔管缺陷检测、残次品剔除与自动分拣。采用缺陷自动检测算法,利用计算机视觉检测技术进行缺陷边缘检测,分割出笔管的缺陷区域并定义主要缺陷类型,完成对笔管缺陷的判断与分类。通过构建、训练卷积神经网络,得到了拟合度较高的卷积神经网络模型,用于分析笔管的缺陷情况。实验结果表明,笔管缺陷自动化检测系统可以客观地检测笔管的缺陷,提高笔管生产效率,提升生产线的成品质量,具有较高的工程应用价值。

Design and research of pen tube defect automation detection system

China has become a major producer and exporter of pen products in the world. One of the key technologies in pen making industry is pen tube testing. Aiming at the requirement of automatic inspection, a pen tube defect automation inspection system was designed to improve the pen tube defect detection efficiency and production quality of pen tube manufacturing enterprises. Based on the machine vision and the gravity center classification device, the defect forms and types of pen tubes were identified and counted by modular detection system to achieve high-efficiency and high-precision pen tube defect detection, defect elimination and automatic sorting. Using the automatic defect detection algorithm and computer vision detection technology, the defect edge detection was carried out, the defect area of pen tube was segmented, the main defect types were defined, and the pen tube defects were judged and classified. By constructing and training convolution neural network, a convolution neural network model with high fitting degree was obtained to analyze the defect of pen tube. The experimental results show that the pen tube defect automation detection system can objectively evaluate the defects of pen tube, raise the production efficiency of pen tube, improve the quality of finished products of the production line, which has higher engineering application value.