GRAIN BOUNDARY DETECTION USING COMPUTER-ASSISTED IMAGE PROCESSING

The properties of rock materials are influenced by the mineral composition, texture(grain size and shape), fabric. A number of petrographic techniques to quantify the mineralogical and textural characteristics have been developed using optical microscope. In petrography, image processing techniques can be applied to examine those characteristics and extract the features of interest from the microscope image of rock thin sections. The identification of grain boundaries using computer-assisted image processing is possible using a quantitative measurement of rock texture such as mineralogical composition, mineral grain size, shape and bonding characteristics of individual grains. However, automated determination of grain boundaries often leads to inaccurate results due to the nature of petrographic image under microscope which makes it difficult to distinguish grain boundaries between two adjacent grains. In the present study, a new method using the Photoshop program has been developed to detect clear and exact grain boundaries between individual grains, even though it is a little time consuming than the automated methods. This method can be applied to evaluate the rock properties, including the fracture toughness of rock materials, from the results of petrographic characteristics measured by image analysis.