Designing And Researching Of Medical Bottle Cap Quality Inspection System Based On Machine Vision

In the production of medical aluminum-plastic bottle caps,there would be various kinds of defective products,such as oil stains,gaps,abnormal roundness,abnormal height and so on.Such defective products flow into the market.It will be affecting the corporate image and the health of drug users.If keep relying on manual testing,it would not only have low accuracy,but also have high labor costs.In view of this situation,this paper is designing and developing a high-speed intelligent detection system for medical bottle caps based on machine vision.At present,the system has been applied in practice.Firstly,the transmission system and image acquisition system will be introduced.The system adopts a two-stage conveying structure,with special structure such as end-to-start overlap and vacuum adsorption,so that the bottle cap runs stably on the belt,and the bottle cap is turned over to take the phase and comprehensively detect.The image acquisition system uses six high-speed industrial cameras,one on the front and one on the back of the bottle cap,and four on the side of the bottle cap.Secondly,it introduced the realization of the detection algorithm.Aiming at the image characteristics of the side,front and back,a specific detection algorithm was designed to identify various types of defective caps.In the image preprocessing and segmentation positioning,the commonly used image enhancement,image segmentation and image matching algorithms are introduced,and the characteristics of various algorithms are analyzed.According to the specific situation and the image characteristics of the bottle cap,an image positioning method based on the idea of regional derivation is proposed.In the detection algorithm,dirt detection,roundness and height detection,six-bridge detection,color mixing detection and other algorithms are designed respectively to realize the detection of various defects.Finally,the mechanical structure and algorithm are tested.After a large number of experiments and field tests,the stability and reliability of the system were verified.The detection speed was up to 950/min and the rejection accuracy was 99.9%.In general,the system has high detection accuracy and high automation,which meets the testing requirements of cap manufacturing enterprises.So it's capable of replacing the traditional manual detection method,improving the enterprise production efficiency,and has broad application prospects.