四川轻化工大学学报（自然科学版）2024，Vol.37Issue(3)：34-41,8.DOI:10.11863/j.suse.2024.03.05
袋式输液器液袋组件管路缺失视觉检测研究
Study on Visual Detection of Missing Pipeline of Liquid Bag Assembly of Bag Infusion Device
摘要
Abstract
Manual quality inspection of the liquid bag assembly will bring quality inspection risk due to subjectivity and fatigue and also reduce the automation degree of the production line.To solve this problem,the machine vision system is adopted to inspect the assembly quality of the lanyard and dosing two passes of liquid bag components.The image acquisition assisting mechanism is designed to solve the problem of difficult image acquisition due to the unfixed shape of the liquid bag assembly;based on determining the defect detection strategy,the Canny algorithm is selected to extract the clear contour of the liquid bag pipeline area in the shortest time,which saves 160～520 ms compared with other algorithms;using the virtual straight-line method,combined with the region of interest(ROI)selected by this paper based on the location characteristics of the outer rectangular box of the contour,the statistics of the number of contours are completed,and the defect detection goal is finally realized.Experiments show that the pipeline defect detection rate of 100%,the maximum detection time of 438 ms,and the detection of a group of liquid bag components need only 1.75 s,much smaller than the factory production line production of a group of liquid bag components of a single station working time of 8 s.The method can replace the existing manual visual inspection,which reduces the cost of enterprise employment.关键词
机器视觉/袋式输液器/缺陷检测/图像处理/边缘检测Key words
machine vision/bag infusion device/defect detection/image processing/edge detection分类
信息技术与安全科学引用本文复制引用
黄波,刘香,张骞,刘康,刘建宏..袋式输液器液袋组件管路缺失视觉检测研究[J].四川轻化工大学学报（自然科学版）,2024,37(3):34-41,8.基金项目
四川省科学技术厅项目(2021YFG0050) （2021YFG0050）
人工智能四川省重点实验室项目(2020RYY01) （2020RYY01）
