Overview
DeepBead supports the on-machine integration of video-based deep learning models for defect detection during the additive manufacturing of metal matrix composites (MMCs). It can be deployed on industrial machines to support both R&D activities and production by providing near real-time analysis of monitoring data. The current version supports the integration of up to two monitoring systems and the detection of up to six different co-existing anomalies in the microstructure of MMCs. To enhance the analysis of monitoring data, predictions from individual video models can be fused for more effective defect detection.
DeepBead provides the following functions:
Features
04- 01
Selection of different types of machine learning models and monitoring systems.
- 02
Integration with optical monitoring systems to support on-machine process parameter qualification.
- 03
Fusion of data from different monitoring systems to enhance machine learning-based predictions.
- 04
Visualization of process parameters, machine learning inferences, monitoring feeds, microstructural defect analysis, and recommendations.