Application of computer images processing technology to analyze materials microstructural images, particularly metallographic images, has received increasing attention. The metallographic images contain the mesoscopic information on structural relation and components of materials. Quantitative analysis of these images can help to correlate the materials structures to their performance and properties at various levels. There are two challengeable issues necessary to be resolved, i.e., automatic segmentation and classification of different microscopic structures in metallographic images. Since the metallographic images often contain complex textures, the segmentation of them is usually inaccurate with present methods. We propose a hybrid algorithm, which combines the Gaussian filter, the mean shift method, the FloodFill, the improved flow-based difference-of-Gaussians, and the clustering to resolve the issues. The experiment results and the comparative results show that our method is effective to segment and classify the microstructural elements in metallographic images with complex textures.
Recently, techniques for nanoparticles have rapidly been developed for various fields, such as material science, medical, and biology. In particular, methods of image processing have widely been used to automatically analyze nanoparticles. A technique to automatically segment overlapping nanoparticles with image processing and machine learning is proposed. Here, two tasks are necessary: elimination of image noises and action of the overlapping shapes. For the first task, mean square error and the seed fill algorithm are adopted to remove noises and improve the quality of the original image. For the second task, four steps are needed to segment the overlapping nanoparticles. First, possibility split lines are obtained by connecting the high curvature pixels on the contours. Second, the candidate split lines are classified with a machine learning algorithm. Third, the overlapping regions are detected with the method of density-based spatial clustering of applications with noise (DBSCAN). Finally, the best split lines are selected with a constrained minimum value. We give some experimental examples and compare our technique with two other methods. The results can show the effectiveness of the proposed technique.