Study On The System Of Filling Quality Monitoring Of Beer Using Machine Vision

At the present time,the brewery detects the sealed lid of the beer bottle, liquid level and impurity or crackle by the eye in china.This method has quite a few disadvantages,such as the lower speed,the low level of efficiency,the high level of omission factor,the workers of detection fatigue easily which will result in a reduction of quality.The machine vision technical system can achieve the high speed,the high precision and automated inspection.The research on vision-based beer machine in our country lagged behind developed countries.We were in urgent need of filling up the gap to develop our beer industry extensively.So this thesis makes an intensive study of the machine vision technical system.The thesis introduces the concept of the beer filling detection,market demand and related important machine vision technology.And designs the machine vision software system.And researches the multiple vision detection methods of the sealed cap of the beer,liquid level and impurity or crackle in great depth.The content as follows:1.Design automatic inspection system in filling beer which is based on an automatic inspection technology and relevant research from home and abroad,combined the effective demand of beer inspection.2.The thesis introduces the concept of the beer filling detection,market demand and related important machine vision technology,includes light source and lighting techniques,camera lens and high speed CCD camera, image grabbing card and the vision sensor and so on,in addition it contains the machine vision is widely used in condition monitoring,inspection and quality control of products in the field.3.On the base of analyzing the inspection demand of the sealed lid,it gave opinion about the shape matching detection algorithm and discussed advantages and disadvantages of the various edge detection algorithm.On the base of analyzing the demand of liquid level,it adopted the fast liquid position measurement in specific area based on the edge detection algorithm. The success rate surpassed 90%,the detection will be done within 15 ms.The experiments showed the algorithm has more perfect practical value.4.On the base of analyzing the inspection demand of impurity and crackle,first traditional edge detection algorithm was proposed,but it couldn't totally satisfy with detection demand.And then the difference shadow detection algorithm was proposed by the experimental analysis,but to no effect.Finally,neural network was used to judge inspection of impurity or crackle for defect,which tested the BP neural network application effect. The success rate of this method surpassed 90%with ideal effect.Theoretical analysis and plentiful experimental proved the rationality of the beer automatic inspection system,the validity and feasibility of the sealed lid of the beer bottle,liquid level and impurity or crackle detection method, which settled most of the troubles in practical application,and laid a good foundation for further research.This system can finish the sealed lid of the beer bottle,liquid level and impurity or crackle detection,and also can select off-test product.