Image Analysis of Meltpools in Additively Manufactured Artifacts
When
Where
Additive Manufacturing (AM) is a sustainable alternative to conventional manufacturing with additional benefits of complex and optimized design. My talk will be a deep dive into the process of analyzing the quality of an AM artifact, specifically into the applications of machine learning (ML) within the analysis process. In order to analyze the quality of the artifact, meltpool analysis is done on a cross-section of the artifact. However, the bottleneck to analysis is segmentation of the individual meltpools. Rather than manually segmenting millions of meltpools, a pixel-based ML algorithm was used to segment the image, with minimal manual corrections. Results of this ML segmentation will be discussed as well as the beginnings of meltpool analysis.