| The ornamental fish has a beautiful shape and bright colors,and is deeply loved by people.Its market potential is increasing and it has become one of the most important aquatic products in our country.Appraisal of the appearance quality of ornamental fish is a major problem in the ornamental fish market.Blood parrot fish is an important species of ornamental fish in our country.Its shape is an important factor in grading.The quality of grading directly determines its market value.At present,the classification of blood parrot fish mainly relies on manual work,and there are problems such as low efficiency and subjectivity.In response to these problems,this thesis uses machine vision technology,based on national standards,to study and establish a blood parrot fish shape evaluation model,which mainly includes two sub-models: body type evaluation model and body color evaluation model.The main work and results are as follows:(1)An image acquisition system of blood parrot fish based on machine vision was designed.The system can quickly capture moving objects and automatically save images to the specified folder.This paper introduces the image preprocessing algorithm of blood parrot fish,and preprocesses the original image of blood parrot fish by means of image enhancement,gray scale transformation and threshold segmentation.(2)Three body size evaluation models including body size classification model,body length/body height ratio model and fish crown model were established.The method of using metric function to calculate the approximate degree of the fish contour and the standard circle and to establish the grading model of the fish shape was emphatically analyzed.The body length and body height of the blood parrotfish were obtained by the long and short axes of the processed image,and a classification model was established by using the body length/body height ratio and a classification model was established by judging whether the blood parrotfish had a crown.(3)The body color evaluation model of blood parrot fish was studied and established.This paper focuses on the extraction method of the body color of blood parrotfish,and determines the average body color of blood parrotfish through regional segmentation,average pixel value,Euclidean distance calculation and other steps,and compares it with the standard color of blood parrotfish,and completes the classification according to the body color.Based on this model,the body color of blood parrot fish is used to identify.After verification,the classification accuracy of blood parrot fish by this method reached 95.5%.(4)This paper designs the shape evaluation software of blood parrot fish based on machine vision,and two sub-models of body size evaluation model and body color evaluation model.The test shows that the method and model of this thesis can comprehensively evaluate the body shape and color characteristics of blood parrot fish,realizes the automatic evaluation and grading of the shape quality of blood parrot fish,and improves the automation and intelligence level of blood parrot fish shape evaluation,solves the problems of traditional methods such as low efficiency and strong subjectivity. |