In-Situ Optical Tomography for Selective Laser Melting

Selective laser melting (SLM) has become one of the most common metal additive manufacturing (AM) processes in industries such as medical and aerospace due to its ability to produce highly specialized end-use parts of great complexity. However, current SLM products are prone to process-induced defects, most importantly porosity, which can greatly alter the strength of the part. Therefore, in this study, in-situ monitoring is investigated as a method for detecting the formation of defects during the SLM process. Layerwise optical tomography (OT) images are collected during SLM fabrication of six test samples with planned defects, and subsequently undergo image processing and analysis for identification of anomalous signatures during the build process. The image processing framework employs an average squared difference (ASD) metric to elucidate defects within the image, then feeds the output into an automated k-means clustering algorithm for segmentation and classification. While the k-means classification approach ultimately proved to be sensitive to noise in the images, the image processing workflow was still able to isolate defects, often providing a clear segmentation of the defect in the final output.