Noise Removal and Image Segmentation in Micrographs of Ferrite-Martensite Dual-Phase Steel

Abstract

 Present work proposes a new methodology for removal of noise (artifacts) from the micrograph, which normally poses a problem during segmentation of phases and measurement of their volume percentages. Segmentation of the phases is one of the primary steps towards quantification of microstructures. Dual-phase steel micrographs, generated from light microscope, consisting of ferrite (white) and martensite (black) phases are segmented by implementation of Otsu threshold algorithm, a well-known threshold algorithm of digital image processing. After thresholding, the noise removing algorithm is applied in different modes. The obtained results are compared with the findings of a commercial metallographic image processing software (Axio Vision). The proposed scheme is found to segment ferrite-martensite dual-phase micrographs successfully.