Research On Contour Recognition And Determination Of Brine Point Position Of Irregular Shape Object Based On Machine Vision

Add bittern into betel-nut is an important link in the process of betel nut processing.Because the appearance of the cut betel nut is irregular,it is difficult to halogen by machine.Therefore,most of the halogen operations of betel nut enterprises are mainly manual halogen,which has the problems of high cost,low efficiency and difficult quality assurance.In order to realize the automatic operation of adding bittern,this paper takes the betel nut which needs adding bittern after cutting as the research object,and combines the machine vision system with motion control to design a visual marinating system,so as to improve the production efficiency and accuracy of marinating.The vision system is composed of two modules: the vision system and the motion control system.The vision system is mainly used to recognize the betel nut on the betel nut pendulum plate and obtain the contour feature information and the point of the betel nut image.The motion control system is mainly used to complete the operation of the betel nut.In this paper,the main research content is the visual processing system.The function of the visual system is to realize the areca contour recognition and obtain areca point of brine.（1）According to the specific requirements,the overall scheme of the visual point brine system is designed,the hardware required by the visual point brine system is analyzed,the experimental platform of the visual point brine system is built,and the operation process of the visual point brine system is described in detail.（2）Because the camera will produce distortion in the imaging process,it is necessary to calibrate the camera.By analyzing the coordinate system of the camera system,the MATLAB camera calibration toolbox is used for image calibration and distortion correction.According to the actual situation of the visual point machine,the affine transformation is used for system calibration,and the transformation relationship between the image coordinate and the actual coordinate is obtained.（3）The feature extraction of inner contour of betel nut is studied and analyzed.Image preprocessing,threshold segmentation,region growing method,image morphology,image bit operation,Retinex image enhancement and other operations are carried out for the collected betel nut image to be marinated.The basic principles and actual effects of several edge detection methods are analyzed,and one of them is determined as the inner contour of betel nut The realization method of contour extraction.（4）According to the shape of the inner contour of betel nut,the least square ellipse fitting method is studied to obtain the point position of the point brine.Based on the VS platform,a visual point brine software system is designed,and the reliability of the algorithm of contour extraction and point brine position determination proposed in this paper is verified by experiments.