Result: New Method for Microstructure Segmentation and Automatic Grain Size Determination Using Computer Vision Technology during the Hot Deformation of an Al-Zn-Mg Powder Metallurgy Alloy.

The microstructure plays a crucial role in material properties, with grain size analysis being essential. Thus, this study introduces a novel approach that utilizes the OpenCV, SciPy, and NumPy libraries in Python, enabling efficient microstructure segmentation and automatic determination of the ASTM grain size number, while clearly defining grains, pores, and boundary regions. In our experiments, an Al-Zn-Mg alloy examined, which underwent deformation at different temperatures (300, 400, and 500 °C) and various strain levels, all at a constant strain rate. Firstly, convert red, green, and blue (RGB) images to grayscale and then, apply median blur to smooth them. The Otsu method was then used to distinguish grains from boundaries using thresholding. This yielded a binary image with differentiated grains and boundaries. Then, further divided the grains accurately using erosion and dilation filters. The binary image underwent additional processing to eliminate noise and classify it into three categories: grains, grain boundaries, and pores. Connected components analysis was employed to identify and label distinct regions in the image, helping determine the number of grains present. By comparing the automated counting method to manual counting, an average relative error of 3.07% was achieved for grain count validation. Furthermore, ASTM grain sizes were calculated based on the number of grains in the optical images, resulting in a high success rate of 99%. These results highlight the effectiveness of the approach in accurately characterizing microstructures and acquiring essential information regarding material properties. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Materials Engineering & Performance is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)