Design And Implementation Of Automatic Counting System Based On Machine Vision

In the edible oil bottle filling production after the end of the oil bottle will be fixed points of the pressure ring and packaging,and due to pressure ring or packaging machinery and other aspects of the problem,there will be missing the number of oil bottles,pressure ring missing,These conditions will bring greater market risk to edible oil companies.Therefore,in the high-speed operation of the pipeline how to use advanced technology to improve efficiency,reduce costs,reduce false positives and missed rate is the key.Based on the analysis of domestic and foreign machine vision technology,this paper puts forward the overall design scheme of automatic inventory system based on the actual demand of oil bottle inventory and ring deletion detection.The paper first introduces the concept of automatic inventory system after filling and the current situation of the urgent needs of the market,as well as some important machine vision technology applications.The hardware part of the system includes light source and lighting technology,optical lens and high-speed camera,the article combined with the characteristics of production lines and imaging requirements given their selection basis.The image processing software of the system has chosen the LabVIEW platform based on G language.The article introduces the function of the platform and the software interface and function of the automatic inventory system.The algorithm design module automatically handles the identification of the number of oil bottles and the defects of the lifting ring through image preprocessing denoising,cap positioning,and ring positioning.Experiments show that the proposed method can effectively improve the detection rate and efficiency,to meet the practical application requirements.The main features of the paper are:●In order to improve the image quality,noise interference in the image preprocessing module,using a median filtering algorithm based on gray value,experiments show that this method is better than the standard median filtering algorithm can not only eliminate the small burr of the oil region reflecting good filtering effect.●In order to determine the position and number of the bottle cap,a Hough gradient circle detection algorithm based on canny operator edge detection is used in the bottle cap positioning module.Compared with ordinary circle detection algorithm,the algorithm can greatly shorten the detection time and improve detection accuracy,and can help solve the problem of sparse distribution produces a lot of noise and the unstable results three accumulator.●Under non-uniform illumination conditions,Niblack’s Method local binarization method detection method was adopted.On the one hand ring box,oil color and body color similar to that of the global binarization method of segmentation is difficult;on the other hand,under non-uniform illumination conditions,local binarization method than global binarization method is applied more widely and more outstanding effect.