Program Development Of Defect Recognition For Three Dimensional Sand Mold Printing

With the transformation and upgrading of traditional foundry industry towards intelligence,informatization and rapidness,3D sand mold printing technology is widely used in foundry industry.In order to ensure uniform discharge and ensure product quality,it is necessary to carry out defect detection on the image of discharge nozzle of 3D sand printing machine.At present,the defects of artificial defect identification,such as low efficiency of defect identification,large number of defects affected by individual subjective factors,and long time of repeated work are not humanized enough.Based on the gradual maturity of machine vision technology in the application of surface defect recognition,this paper USES the WinForm application program in the VisualStudio2015 development tool to develop a set of 3D sand print image defect automatic recognition program to overcome its shortcomings.Based on the understanding of the characteristics of 3D sand print image and the application of machine vision technology in the surface defect identification,the advanta^es and disadvantages of the algorithms adopted by each module were analyzed and compared,and the advantages of different algorithms were combined to make it better applied in the development process of 3D sand print defect identilBcation program.The main research contents and research results are as follows:1.In order to obtain a higher defect identification rate and shorten the ru?ning time of the progra1,this paper first preprocesses the collected 3D sand print defect images.The preprocessing steps are gray-scale by weighted average method,median filtering,binarization by otsu method,mathematical morphology processing and image rotation.2.Use bilinear interpolation method to normalize the size of the 12 regions of interest in the image after rotation correction,so as to facilitate the segmentation of the upper and lower regions of each image of interest.Then,the normalized interested image region is divided into two defect regions by finding the dividing line of the middle region,which provides the feasibility for the further development of the image defect recognition program.3.through the study of the pixels of the defective area of the pixel value is 0(black),to eliminate the defect area and keep defect area(white area of pixel value of 255),and then the regional fluctuation of the separated package edge processing,make the image of the black and white form a closed in the region and to facilitate the final number of 3D sand mold to print the image defects identification work.4.Get the minimum inner rectangle of the white contour area,set the rectangle with the width greater than 4 pixels as the defect,screen the defect area,and finally display the number of defects in each area identified on the computer screen according to the original image serial number.Through a large number of experimental results,it can be seen that the developed system program can identify defects accurately and quickly.In this paper,the average recognition rate of defects in the region of interest calculated by two methods is 90.82%and 85.55%,respectively.The average recognition time of each image is 13.56s,which can meet the needs of practical application.