Research On Tomato Quality Inspection And Classification Based On Machine Vision

Detection and classification is an important part of the fruit and vegetable grading.The sorting of tomatoes is mainly relying on manpower in the standardization of fruit and vegetable grading in my country,and the disadvantage is time consuming,the efficiency is low,and the detection and judgment of people affected by subjective and objective factors.Error,which brought uncommon difficulties to grade processing,grading,hierarchical exit,etc.Based on the lack of tomato grading mode,this paper studies the machine-based tomato quality testing and classification study on the quality of the tomato and calculates the maturity,the size,shape,shape,and calculates the score,through comprehensive evaluation of three key factors.It has been obtained more applicable tomato quality grading model,thereby better improving detection and classification efficiency,and the specific research content is as follows:(1)Investigate domestic and foreign scholars about the design of tomato detection technology,this experiment determines the installation position and shooting angle of the camera,equipped with a suitable light source and a small white grinding bottom plate to ensure high image,maturity analysis Tomato horizontal bilateral images and synthesize an image for data processing,diameter detection and shape detection by acquiring a top view of the same data for data processing.(2)During the image pretreatment,grayscale processing is first performed,and the Gaussian filtering method is selected as the image filtering method by testing comparison,and is used to remove the noise of data.Conversion of RGB color space to HSV color space to study color image color characteristics,maturity calculation,extracting H component custom threshold is(0-8,60-255,60-255)∪(160-180,60-255)60-255)Make image segmentation.The OTSU Essence is used as a contour characteristic algorithm for tomato diameter detection and shape detection.After the image preparation,the morphological processing is performed,and the opening operation of the nuclear value size 5X5 is an iterative six times for processing effects.(3)Collect 500 sheets of tomato data,Python write code,analyze the method of pattern,diameter,and shape.After extracting the threshold interval,first calculate the ratio of the ratio of the inventory of tomato color to the contour area,and then use the minimum external rectangle to realize the diameter of the tomato size.Finally,the outer cutting circulation is calculated by the minimum outer cutting process.The ratio determines the degree of rule of the shape.This topic has established a set of tomato quality calculation score algorithm,plus 0.3 ratio of 0.3 ratio of diameter size and 0.2 ratio of the shape of the diameter size as the final quality grading model with0.5 ratios of maturity.(4)After establishing a good tomato quality grading test model,this topic divided four levels,which were S-class tomatoes at 0.9 points or more,and the quality score was at a level tomato,quality division.A unified classification of the value of B-grade tomato and a score of 0.7-0.8 is less than 0.7 or less.By creating the sample library,the set interval period is extracted from a certain number of tomato samples from the sample library to compare,and the accuracy of the final overall result is 90% by comparing the model classification algorithm designed by the artificial grading method.above.