Design Of Aquatic Product Counting And Grading System

This paper designs a system of counting and classifying aquatic products.At present,the main method of classifying aquatic products in China is artificial.Another method of counting,classification and classification is to use machine vision detection technology,and use machines instead of human eyes for detection.In machine vision,costs can be reduced because the system hardware is easy to integrate and implement.It has the advantages of fast classification speed,high classification accuracy,simple use and convenient maintenance,which is not possessed by traditional classification technology.Therefore,it is necessary to speed up the efficiency of counting,classification and classification of aquatic products.The system of counting and classification of aquatic products is designed by using machine vision technology.In order to solve the problem of the number and classification of aquatic products,this paper first collects the image of aquatic products.Two methods are used to process the image.One is based on OpenCV and USES the minimum external rectangular area to classify and count aquatic products.The second classification method is based on the multi-layer perceptron to classify aquatic products.The experiment proves that the machine vision technology has a good effect on the classification,classification and counting of aquatic products.The specific research contents are as follows:(1)Research and select the industrial camera,lens and light source of the hardware system,and finally select the most appropriate hardware system based on the actual environment and scientific research funds.(2)Using OpenCV and MFC mix of counting and classification of aquatic products,based on the area of the minimum circumscribed rectangle for counting and classification of aquatic products,mainly including the image gray processing,image filter processing of aquatic products,such as binarization image preprocessing,find the image of the minimum circumscribed rectangle of aquatic products,and its computing area,etc.,among them,the image filtering are introduced bilateral filtering,median filtering,gaussian filtering,after comparing several kinds of filtering effect of the image processing of aquatic products;The effects of single thresholding and adaptive thresholding on aquatic product images in binarization were also studied,and the appropriate thresholding was selected through comparison.(3)Halcon and C# are used to realize the counting and classification of aquatic products.The counting and classification of aquatic products are carried out based on the method of multi-layer perceptron.The algorithm of multi-layer perceptron,the creation of multi-layer perceptron,the training of multi-layer perceptron and the online classification of aquatic products are mainly studied.(4)Encapsulate the codes of OpenCV and halcon,and visualize them using MFC and C#respectively to realize the operation of the system.It can count and classify aquatic products,and the operation result basically meets the expected effect.The two methods are compared to select the optimal design scheme.The experimental data show that the system can effectively classify and count aquatic products.