Segmentation of dual phase steel micrograph: An automated approach

Digital image processing is used to analyze the microscopic images of the materials. Extraction of grains/phases present in the material is the fundamental step to achieve the description of the microstructure. In this work, we present an automated scheme for segmenting the phases present in the scanning electron microscopic images of dual phase steel. The challenges posed by the presence of revealed grain boundaries bearing striking similarity with one of the phases and intensity variation within the phase regions also has been addressed successfully. The proposed scheme successfully approximates the closed phase regions for a variety of micrographs. Moreover, the proposed scheme does not make any assumption regarding the factors like magnification and any other imaging condition affecting the image characteristics.