Identification of transcrystalline microcracks observed in microscope images of a dolomite structure using image analysis methods based on linear structuring element processing

This paper presents an image analysis algorithm for transcrystalline microcracks recognition. The algorithm is applied for the analysis of microscope images of a dolomite structure. Linear structures are derived using an image analysis method based on mathematical morphology functions utilizing linear structuring elements. Next, the standard deviation function of microcracks and grain boundaries of surrounding pixel values is used to determine transcrystalline microcracks. All operations were executed in CIELab colour representations of dolomite images. The results obtained confirm the adequacy of this approach as a robust segmentation tool. Accordingly, an algorithm was derived that allows a fully automatic segmentation of transcrystalline microcracks. This is followed by geometrical and statistical analysis. The developed algorithm may facilitate petrographical and stereological studies of rock structures observed under a polarizing microscope. (c) 2006 Elsevier Ltd. All rights reserved.