Research On Defect Inspection Method Of Sintered Abrasive Blocks Based On Machine Vision

Barrel finishing is a kind of widely used machining technology to improve the surface quality of the parts,when processing,putting parts and abrasive blocks,water,and abrasive fluid into the drum,through the roller rotation made roll abrasive blocks and parts generate collision,friction and rolling force,so as to remove the burr,scratch on the surface of the parts,to improve the surface brightness and surface quality of parts.Abrasive block is the abrasive medium in roller mill processing,which has an important effect on the machining effect of parts.Sintered spherical abrasive block is one of the most widely used in surface finishing of parts.But at present,due to the factors of production technology and equipment,the sintered spherical abrasive blocks produced by domestic manufacturers usually have some defects such as abnormal roundness and black core.The existence of these abrasive block defects will seriously affect the service life of the abrasive blocks,and even affect the surface quality of parts.However,at present,the domestic defect detection of abrasive blocks mainly relies on manual judgment,which has some problems such as strong subjectivity,low detection efficiency and inconsistent detection standards.In order to more efficiently and accurately detect the abnormal roundness and black defects of sintered spherical abrasive blocks,this paper proposes a defect detection method of abrasive blocks based on machine vision,which mainly includes the following research contents:(1)Designed and developed independently a set of abrasive block image automatic acquisition system,including: sampling disc,single-chip microcomputer,stepping motor,digital microscope and upper computer.The stepper motor is controlled by a single chip microcomputer,and the stepper motor drives the sampling disc to rotate,and the abrasive block is rotated to the lower part of the digital microscope lens.Through the signal transmission of the single chip microcomputer and the upper computer,the rotation of the stepping motor and the time sequence of the digital microscope can be controlled,so that the system can coordinate and complete the automatic acquisition of abrasive block image.(2)In order to improve the segmentation quality of abrasive block image,the grayscale algorithm of abrasive block image is studied,compared and improved.Firstly,the grayscale image of the grinding block is grayscale by the maximum value method,average value method and weighted average value method.The grayscale image of the grinding block is evaluated by the image evaluation indexes such as standard deviation,average gradient and information entropy.Secondly,the gray-scale method of component ratio is proposed for the abrasive block image with poor gray-scale effect.Finally,the abrasive block grayscale image is effectively segmented by OSTU threshold segmentation algorithm.(3)The noise points in the image of abrasive block are analyzed and removed,and the defect size of abrasive block is extracted and calculated.The hole-filling algorithm,image opening operation and median filtering were used to remove the small holes,isolated noise points and smooth the edge of the abrasive block image.The area of the abrasive block and the black defect area are extracted by color filling,and the area of the abrasive block area,the circumference of the abrasive block contour and the black defect area of the grinding block were calculated.The roundness of the abrasive block is calculated by the roundness calculation formula,and the proportion of the black core defects of the abrasive block is calculated according to the area of the black core defects and the area of the abrasive block.Finally,the defect size distribution range of qualified abrasive blocks and unqualified abrasive blocks is analyzed and determined through the abrasive block defect detection experiment.The experimental results verify the effectiveness of the abrasive block defect detection method based on machine vision.(4)Using C# language developed abrasive block defect detection PC interface,including abrasive block image acquisition interface,the interface of image processing and abrasive block image management interface,can be collected in real-time display digital microscope abrasive block image,image processing and abrasive block defect detection results,as well as the realization of abrasive block images and abrasive block defect information management.