Automatic characterization of spherical metal powders by microscope image analysis: a parallel computing approach

Abstract The microstructural characteristics of spherical metal powders play an important role in determining the quality mechanical parts manufactured by powder metallurgy processes. Identifying individual particles from their microscopic images is one most convenient and cost-efficient methods for analysis characteristics. Although numerous image processing algorithms have been developed automating particle identification process, they perform less accurately identifying adjacent that are heavily overlapped regions. We propose automatic algorithm to robustly identify particles, especially microscope images. A parallel computing framework designed further enhance computational efficiency proposed algorithm. Powder such as size distribution location potential satellite derived our results. accuracy validated real-world scanning electron images, outperforming other existing achieving both precision recall above 99%.