Research On Liquid Liquor Impurity Detection System Based On Machine Vision

As an important part of Chinese culture,wine plays an irreplaceable role in people's social life.In the process of liquor production,there are many production links and complex processes,especially in the filling link,because The filling bottle is not cleaned,the filtration system is faulty,and the gland packaging machinery collides,causing impurities such as glass fragments,flocs,hair and the like to be mixed into the liquid in the bottle.If these liquors containing impurities flow into the market,not only does it seriously damage consumers' health.But also it will have a tremendous impact on enterprises and society.Aiming at the inefficiency of liquid impurity detection methods for transparent bottled liquor and the complexity,poor anti-interference ability and high price of existing impurity detection systems,an online detection method based on machine vision is proposed.Based on this method,a detection system for liquid impurity of transparent bottled liquor is designed,which is mainly based on the motion characteristics of the impurities in transparent bottled liquor.According to the movement characteristics of impurities in the transparent bottled liquor,the transparent bottle image is continuously collected and analyzed by an industrial computer-driven industrial camera to detect impurities in the liquid.Firstly,according to the proposed visual detection method and design requirements,the detection platform is designed,including image acquisition equipment,light source and lighting equipment,system mechanical hardware and software platform.The image acquisition equipment adopts MV-GE1000M-T CMOS industrial camera of Medvedev Company.The light source and illumination mode adopts parallel LED backlight.The mechanical hardware of the system is mainly vertical flip mechanism.The software platform of the system is Windows platform based on Visual Studio 2013 of Microsoft Company and Halcon of machine vision image processing algorithm library of MVTech Company of Germany.According to the characteristics of impurities in liquor,the design of image processing detection algorithm is the main content of this paper.In this paper,the basic concepts and methods of image processing technology are introduced,and image filtering,edge detection,gray transformation,feature extraction,inter-frame difference method,KNN learning algorithm are combined to realize the design of liquid impurity detection algorithm for transparent bottled liquor,which can distinguish bubbles,stains and impurities in transparent bottles with dynamic interference.In this paper,the upper computer of the detection system is designed by programming,which mainly includes two parts: background program logic design and human-computer interaction interface.The host computer uses C# language and Halcon image processing algorithm to realize the detection and removal of impurities in liquor on Windows platform.This paper validates and analyses the designed detection system by setting up an experimental platform.The experimental results show that the real-time performance and stability of the proposed liquor impurity detection system have been greatly improved even under dynamic interference environment.To a certain extent,the problem of impurity detection in the automatic production process of transparent bottled liquor has been solved.