Study On Onlin Inspection And Classification Of Potato Quality Based On Machine Vision

In recent years,machine vision technology has developed rapidly and has been gradually applied to potato quality detection and classification,potato quality classification based on machine vision can effectively avoid secondary damage caused by mechanical detection,and can also exclude the interference of subjective factors on classification,ensure stable work for a long time,improve detection efficiency and classification accuracy.This paper is based on machine vision to study potato quality detection and classification,and the main research are as following:Select the camera lens,and use the Zhang Zhengyou calibration method to calibrate the camera to complete the image correction.Preprocess the original potato image with grayscale,filtering,threshold segmentation,Through comparison,choose the best pretreatment method.In the part of detection and classification of potato external defects,select the method based on the Euclidean distance of RGB space for green skin detection by analyzing the characteristics of external defects such as green skin,dry rot,holes,sprouting,mechanical damage,use SUSAN operator to detect other defects except green skin.The two methods are combined to determine the detection process of external defects and remove the defective potatoes.In the part of detection and classification of potato shape,invariant moments are selected as the characteristic parameters of shape detection.On the basis of the existing seven invariant moments,three additional invariant moments are added through the trigonometric function generation method to make the detection results more accurate,the ten invariant moments of the potato edge detection image are input to the BP neural network for training,and the trained neural network model divides the potato into two types: non-malformed and malformed.Select the kinect v2 camera to detect and grade the size of the potatoes.Through the principle of time of flight,the potato depth image and its background image can be obtained to obtain the surface thickness image.The pixel by pixel integration method is used to obtain the potato volume and then obtain the potato weight.According to theweight,the potato is divided into three levels: large,medium and small,and an image processing system for potato quality detection and grading is built.