A Study Of Nodular Cast Iron Microstructure Segmentation Based On The Image Processing Technique

Traditional metallographic image organization analysis mostly adopts the way of combining the human eyes with personal experience to judge and classify the metallographic microstructure. The results of analysis are often influenced by subjective factors. Labor intensity is high, and work efficiency is low. To a certain extent, these lead to lack of the accuracy and objectivity of related data. Number9nodular cast iron microstructure is segmented by modern image processing technology. Using the image processing related algorithms studies the segmentation of number9nodular cast iron metallographic image. In finally, the better results of segmentation and quantitative statistics are achieved.Metallographic specimens are prepared in accordance with the relevant requirements after sampling, grinding, polishing, etching, cutting and other process. The metallographic images are observed and collected by using computer and microscope. Taking into consideration the characteristics of number9nodular cast iron metallographic microstructure image, the image of process and segmentation adopts the modern image processing related processing algorithms. The concrete research content is as follows:(1) The basic methods of the modern image processing technology is used to process number9nodular cast iron metallographic microstructure image. It can get a better effect. Pre-process removes the noises in the images of the metallographic microstructure and retains the useful detail information.(2) The circular degree detection, Hough transform circle detection and fitting circle detection are used to detect and segment spheroidal graphite of number9nodular cast iron metallographic microstructure image. And the results of these three methods make the comparison to obtain the optimal results.(3) The segmentation of number9nodular cast iron metallographic microstructure image gets the three region areas parameters that are spheroidal graphite, ferrite and pearlite.(4) According to the modern image processing methods, software system integration based on Matlab to form an interactive interface, implement related steps.