Research On Metallographic Structure Analysis Technology Of Cast Iron Based On Image Processing

With the rapid development of manufacturing industry,the demand for metal materials is increasing,and the quality requirements are getting higher.Therefore,it is very important to test the properties of metal materials.The traditional metallographic analysis method has low efficiency,high work intensity and greatly influenced by artificial factors.In addition,the existing computer aided metallographic analysis system still has many shortcomings,such as lack of classification and low accuracy of composition analysis.This paper takes cast iron material as the research object,which combines digital image processing technology and machine learning algorithm to carry out in-depth research on metallographic structure analysis technology.The main contents are as follows:(1)Researching on the acquisition and preprocessing algorithm of cast iron metallographic images.The cast iron metallographic samples were prepared according to relevant standards,and the cast iron metallographic images were collected by metallographic microscope,and the images were preprocessed.After the steps of image cutting,screening and label setting,the metallographic image database of cast iron was established.To solve the problem of noise in the image,the bilateral filter was used to remove the noise without destroying boundaries.An adaptive histogram equalization method with limited contrast is used to solve the problem of uneven brightness in images.It provides an effective technical path for image segmentation and feature extraction.(2)Researching on automatic classification and recognition technology of cast iron Metallographic structure based on texture feature and machine learning.In feature extraction,the improved local binary mode(C-LBP)and gray level co-occurrence matrix(GLCM)were used to extract texture features from metallographic images,and the different extracted features were sequentially fused to analyze the recognition effect of single feature and fusion feature on the metallographic structure of cast iron.In terms of classifier selection,the influence of different kernel functions on the generalization performance of support vector machine(SVM)were compared,and the parameters of SVM model were optimized by the grey wolf intelligent optimization algorithm(GWO).Finally,the recognition effect of different classifiers on the metallographic structure of cast iron were compared.The results show that the recognition rate of SVM model can reach 99.17%,which is better than random forest(RF)and extreme learning machine(ELM).(3)Researching on image segmentation and parameter determination of cast iron metallography.A k-means clustering segmentation method based on Tent chaotic sparrow optimization was proposed,which uses the sparrow optimization algorithm introduced chaos sequence to optimize the initial clustering center of K-means to solves the problem that the kmeans clustering segmentation results are highly dependent on the initial clustering center.The randomness of tent chaos sequence can prevent the algorithm from falling into the dilemma of local optimal solution and achieve convergence more quickly and accurately.The results show that the segmentation accuracy of the proposed algorithm is better than that of the traditional K-means algorithm,and the average running time is 0.3674 s,which can meet the real-time requirements of cast iron metallography analysis.Morphological processing and connected domain analysis were used to solve the interference of noise points in the image and realize the labeling of different target regions.Then feature parameters such as the area ratio,shape factor and minimum enclosing rectangle were extracted.It is proved that the method of metallographic analysis based on digital image processing is more accurate and repeatable than the traditional method.(4)Software design and testing for metallographic analysis of cast iron.The GUI toolbox in MATLAB software was used to complete the development of the software interface of cast iron metallographic testing,to realize the classification and identification of cast iron metallographic structure,the extraction of characteristic parameters of graphite particles,and the use of extracted parameters to grade samples,which can meet the needs of some cast iron metallographic grading.