Design And Application Of Medical Bottle Cap Detection System Based On Machine Vision

As a new type of bottle cap in the pharmaceutical industry,the medical aluminum-plastic composite bottle cap has been widely used due to its advantages of simple structure,environmental protection,and easy opening.But its properties are still detected by artificial visual inspection.Due to the disorientation,off-job,fatigue or lack of sense of responsibility of the testers,obviously,this way has a low efficiency and a high negative detection rate and affects the test result.Therefore,it's necessary to develop an online intelligent automatic detection system with independent intellectual property rights,instead of human eye detection,to achieve higher precision efficiency detection.Most of the cap defects locate on the outer and inner surfaces of the cap,such as the stains on both above mentioned situation,color deviation or incomplete pattern of the trademark logo on the former situation and the incomplete riveting on the latter one.Considering the defects characteristics of the medical bottle cap,firstly,determining the overall design of the system,including the system hardware structure and software architecture,and secondly,determining the hardware according to the system design requirements,hardware structure,and the selection of components,and finally detecting and analyzing the cap through the algorithm.In view of the defects existing on the inner surface of the bottle cap,this study mainly tested stains and riveting.This study detected the integrity and color of trademark pattern,and black heterochromatic spots using Halcon's classic algorithm to perform operations on the image to complete the inspection task.After the bottle cap passed through each workstation,the IPC preprocessed the captured images and used the defect detection algorithm to detect the images.The detection results were displayed in real time.The IPC sent a signal to the data 10 card.This process controlled the air nozzle to shoot the defective bottle cap.The experimental results indicate that the designed algorithm in this study fully meets testing requirements.Finally,the prototype set up was used for real-time online detection of the actual production of medical bottle caps(BC),achieving a target of 1200 BC/min and the test results are given.After 2-month trial operation,the practicability of this research was verified.