| With the application of the digital image technology has been used more and more widely, the quantitative analysis of digital image plays an important part in the metal material research field. In many ways, such as tissue analysis and material checkout, the digital image technology has taken the place of manual work. In the ultrafine grain steel research field, analysis the metallurgical structure by digital image technology is very important to new metal material development.The topic of this thesis is about the metallurgical structure. For the reason of sample preparation and image grabbing technology, grain boundaries in the metallurgical structure images may be fuzzed or disconnected, it has a bad effect on the following quantitative analysis such as grain size checkout. In the past, this work was done by manual work, it was inevitable that errors were produced and more time was taken.In order to solve all kinds of problems in the metallurgical structure images, a newly metallurgical structure image restoring method based on water-growing is proposed through a great number of experiments. According to the irregular shape characteristic of the adhesion grains, the grain boundaries are get by ultimate eroding, numbering the objects, and repeated dilation. So the adhesion grains are split and images are restored. This thesis designs some methods to calculate the metallurgical structure parameter; develops an analyzing and recognizing system of metallographic microscopic structure based on the restoring result of the metallurgical structure images. This system has a visualized interface and human-computer interacting function, automatic measurement of several kinds of characteristic parameters.Experiment results show that, compared with manual restoring results, this method can split grains and restore image very well. It solves the problem in grain boundaries fuzzed or disconnected quickly and conveniently. |
PDF Full Text Request