Design Of Liquid Level Detection System For Propylene Polymerization Process Based On Machine Vision

Polypropylene,abbreviated as PP,is a polymer formed by the polymerization of propylene,which is an important thermoplastic material,as well as an important chemical product and chemical raw material.Global consumption of polypropylene has grown at an average annual rate of 9.5% over the last few decades,making it one of the fastest growing thermoplastic materials consumed.Abnormal liquid levels during propylene polymerization can lead to unsuccessful polymerization and even safety accidents.In order to meet the demand for level detection in the propylene polymerization process,this project uses machine vision and deep learning image processing technology to study and design a machine vision-based level detection system for the propylene polymerization process,and to test the performance of the system.The system consists of a main control unit,a level image acquisition unit,a level image processing unit,a communication unit,and a display unit.The system is based on the Linux operating system and uses the RK3399 chip as the core processor for the system design;the level image acquisition unit uses a binocular camera to capture the level images of the propylene polymerization process;the level image processing unit uses the Mali-T860MP4 graphics processor;the Python programming language is used for programming,and the Py Torch deep learning framework and the results of the level detection and the current reaction status of the reactor are displayed in real time on an LCD unit,and the results are uploaded to the cloud via a Wi Fi communication unit to facilitate real-time monitoring of the propylene polymerization process.Based on a discussion of the current state of research on level detection at home and abroad,this project addresses the shortcomings of traditional level detection systems for level detection in the propylene polymerization process and combines professional knowledge to design an intelligent system for the measurement of level parameters in the propylene polymerization process using machine vision equipped with deep learning image processing technology.Research on target detection and stereo matching algorithms is one of the most important topics in the field of machine vision and image processing,and this project has successfully implemented the YOLOv5 target detection algorithm incorporating the Cost Benefit Analysis Method(CBAM)attention mechanism and comparing different stereo matching algorithms for the measurement of level parameters in the propylene polymerization process.The YOLOv5 target detection algorithm and the comparison of different stereo matching algorithms have successfully achieved the measurement of liquid level parameters in the propylene polymerization process.The application of machine vision technology in liquid level detection is of great importance for the development of liquid level detection technology,as it has the advantages of high real-time and high detection efficiency.In this study,a machine vision-based level detection system for the propylene polymerization process was successfully designed.The accuracy,repeatability and stability of the system have been tested.The accuracy of the system is 1 mm,the absolute error is within 2 mm,the relative error is within ± 5%.The system has the advantages of high detection accuracy,real-time detection and high detection efficiency,which improves the intelligence of liquid level detection systems in the market and has high market application value and good economic benefits.