Synthetic-to-real Composite Semantic Segmentation in Additive Manufacturing

The application of computer vision and machine learning methods in the field of additive manufacturing (AM) for semantic segmentation of the structural elements of 3-D printed products will improve real-time failure analysis systems and can potentially reduce the number of defects by enabling in situ corrections. This work demonstrates the possibilities of using physics-based rendering for labeled image dataset generation, as well as image-to-image translation capabilities to improve the accuracy of real image segmentation for AM systems. Multi-class semantic segmentation experiments were carried out based on the U-Net model and cycle generative adversarial network. The test results demonstrated the capacity of detecting such structural elements of 3-D printed parts as a top layer, infill, shell, and support. A basis for further segmentation system enhancement by utilizing image-to-image style transfer and domain adaptation technologies was also developed. The results indicate that using style transfer as a precursor to domain adaptation can significantly improve real 3-D printing image segmentation in situations where a model trained on synthetic data is the only tool available. The mean intersection over union (mIoU) scores for synthetic test datasets included 94.90% for the entire 3-D printed part, 73.33% for the top layer, 78.93% for the infill, 55.31% for the shell, and 69.45% for supports.

Towards Smart Monitored AM: Open Source in-Situ Layer-wise 3D Printing Image Anomaly Detection Using Histograms of Oriented Gradients and a Physics-Based Rendering Engine

This study presents an open source method for detecting 3D printing anomalies by comparing images of printed layers from a stationary monocular camera with G-code-based reference images of an ideal process generated with Blender, a physics rendering engine. Recognition of visual deviations was accomplished by analyzing the similarity of histograms of oriented gradients (HOG) of local image areas. The developed technique requires preliminary modeling of the working environment to achieve the best match for orientation, color rendering, lighting, and other parameters of the printed part. The output of the algorithm is a level of mismatch between printed and synthetic reference layers. Twelve similarity and distance measures were implemented and compared for their effectiveness at detecting 3D printing errors on six different representative failure types and their control error-free print images. The results show that although Kendall tau, Jaccard, and Sorensen similarities are the most sensitive, Pearson r, Spearman rho, cosine, and Dice similarities produce the more reliable results. This open source method allows the program to notice critical errors in the early stages of their occurrence and either pause manufacturing processes for further investigation by an operator or in the future AI-controlled automatic error correction. The implementation of this novel method does not require preliminary data for training, and the greatest efficiency can be achieved with the mass production of parts by either additive or subtractive manufacturing of the same geometric shape. It can be concluded this open source method is a promising means of enabling smart distributed recycling for additive manufacturing using complex feedstocks as well as other challenging manufacturing environments.