Study On Key Technologies Of Small Citrus Automatic Grading Equipment

Citrus grading is the most important link in the post harvest commercialization of citrus,which can greatly improve the added value of citrus.The expensive large-scale grading line is not suitable for the field grading needs of small and medium-sized orchards.In order to solve this problem,this paper takes the wide peel tangerine as an example,analyzes its characteristics and classification technology,puts forward the automatic classification method,develops the miniaturized automatic classification equipment of citrus,and realizes the automatic classification of citrus.The specific research contents of this paper are as follows:(1)The characteristics of citrus were studied by building a platform test,and its appearance characteristics,motion characteristics and mechanical characteristics were analyzed,so as to provide data support for the research of grading method.The grading rules of Citrus size and defects were studied,and the grading standard suitable for wide skinned citrus was drawn up.In addition,by analyzing the existing manual classification and automatic classification processes of citrus,the key points of realizing classification are summarized,and then the research direction of classification method in this paper is determined.(2)In order to realize the automatic grading of citrus,the sorting method based on Citrus characteristics,the single fruit feeding method based on height difference,the turnover detection method based on friction drive and the trigger grading execution method based on fixed-point conveying are proposed.According to the proposed classification method,the specific structural design of each part is carried out,and the kinematics and mechanics analysis of the key structure are carried out to provide a theoretical basis for the structural design.The curve equation of double cone roller is deduced to realize that the rotation angular velocity of Citrus in the range of fruit diameter is similar on the roller,which is convenient for the camera to collect the whole surface information of citrus;The turnover detection device is simulated and analyzed by ADAMS software to verify the feasibility of the design theory of double cone idler;According to the simulation results,the stability time of overturning motion is determined,which is convenient for the camera to collect images.(3)Citrus specification measurement and defect detection technology based on machine vision provides data support for automatic grading.Based on the analysis of the outer contour characteristics of citrus,a sub-pixel contour fitting measurement method is proposed to measure the fruit diameter of citrus.Aiming at the common scar defects such as black gray and dark brown,the defect feature extraction method is obtained by analyzing their features in HSV color space;The SSD deep learning network is used to distinguish the characteristics of fruit defects,fruit stems and umbilicus,and the defect detection is realized.(4)The feasibility of the classification method and the stability of the mechanism function are verified by making a prototype and carrying out experiments.Taking the artificially graded citrus as the experimental sample,the feasibility and accuracy of the fruit diameter measurement algorithm and defect detection algorithm are verified,and the error is smaller than that of the traditional detection algorithm.According to the classification standard of this paper,the classification experiment is carried out,and the high-precision automatic classification of citrus is realized,which can meet the classification needs of small and medium-sized fruit farmers.