Automated image analysis for quantification of materials microstructure evolution

Abstract In this work, an automated image analysis procedure for the quantification of microstructure evolution during creep is proposed evaluating scanning electron microscopy micrographs a single crystal Ni-based superalloy before and after at 950 °C 350 MPa. microscopy-micrographs ? / ? microstructures are transformed into binary images. Image analysis, which involves pixel by classification feature extraction, then combined with supervised machine learning algorithm to improve binarization quality results. The gray scale images not always straight forward, especially when difference in levels between -channels ?-phase small. To optimize we utilized series bilateral filters as well algorithm, known gradient boosting method, that was used training classifying micrograph pixels. After testing two methods, method identified most effective. Subsequently, Python routine written implemented area fraction channel width. Our documented results automatic discussed based on previously reported literature.