Font Size: a A A

Metallographic Analysis Of Carbides In Steel Based On Digital Image Processing

Posted on:2019-11-12Degree:MasterType:Thesis
Country:ChinaCandidate:Y X JiangFull Text:PDF
GTID:2371330545456480Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
Because the density and intensity of carbide in steel directly affect its hardness,toughness and other performance indicators,the use of metallographic analysis to detect the carbides and grade the steel is an important process in the steel production.However,the traditional metallurgical analyses mainly use artificial visual inspection,manual adjustment,mechanical positioning and so on.The accuracy of the measurement results is greatly influenced by human factors,the precision degree is unstable and the positioning search ability is poor.The digital image processing system is applied to the metallographic analysis,which has the advantages of high precision,high speed and can greatly improve work efficiency.Based on the GB/T18876 "Standard Test Method for Determining Metallography and Content and Grade of Inclusions in Steel and Other Metals by Using Automatic Image Analysis",combining with the theoretical knowledge and tools of metallographic analysis,image segmentation,pattern recognition,mathematical morphology,quantitative metallographic analysis and grading about 06 series banded carbides in steel based on digital image processing are achieved.In the experiment,the image preprocessing is firstly carried out,which mainly includes image type conversion,image denoising and complexity classification based on two-dimensional information entropy.Secondly,in the process of image segmentation,the key of the experiment,the methods of edge segmentation,threshold segmentation and region-based region growth and watershed segmentation are comparatively analyzed.Finally,the multi-threshold segmentation method based on two-dimensional entropy and super-pixel segmentation based on SLIC is combined to finish image segmentation.The partitioned binary image is used to acquire the target areas inside the circle through the fitting circle algorithm.The area of the banded carbides is added up and the area ratio is calculated by mathematical morphological methods,and finally the level of the banded carbide is obtained.At present,this method can accurately classify 06 series banded carbides of all grades and achieve automated grading.
Keywords/Search Tags:Metallographic analysis, Image segmentation, Region growth, SLIC super-pixel segmentation, Mathematical morphology
PDF Full Text Request
Related items