Active Contour Multigrid Model for Segmentation and Automatic Quantification of Material Phases of Cast Iron

Abstract

A critical and important stage in microstructure image analysis is segmentation, because the segmentation method has direct impact on the end results of analysis. The main aim of this paper is to determine appropriate segmentation method for microstructure image analysis and quantification. In this work, some popular segmentation methods, namely, Otsu's automatic threshold, watershed, uni-grid active contour method and multi-grid active contour methods have been investigated. The reliability of the segmentation methods is tested by determining the volume fraction of phases present in microstructure images of materials of known chemical composition. The experimentation is done using microstructure images of cast iron of various compositions. The experimental results are compared with expected values of volume fraction. The active contour multi-grid segmentation model is found to yield better results within the practical limits as compared to manual and other automated methods.

Automatic quantitative analysis of microstructure of ductile cast iron using digital image processing

“…The microstructural images are segmented by Otsu's threshold segmentation technique. Pattan Prakash, V. D. Mytri [6], used Otsu's optimum segmentation method based on thresholding in order to separate nodules. The method does not work well with variable illumination.…”

“…Hence template matching is not suitable option for analysis of nodular cast iron. A series of morphological operations for segmentation of microstructures are effectively used [6], but these methods are dependent on selection of appropriate size of structural element.…”

Online Classification and Measurement of Pencils using Image Processing Techniques

“…Performance analysis of matching and classification of images based on wavelet energy features for object shape recognition from tactile images is presented in [14]. Microstructure image analysis and quantification by investigating the performance of segmentation methods for automatic quantification of material phases of cast iron is reported in [15]. In [16], a technique that employs support vector machines and Gaussian mixture densities to create a generative/discriminative object classification using local image features is presented.…”