[Integrated and lightweight photovoltaics]

Facing the challenges of our time

Predictive Analytics group

• The proposed system will combine computer vision, object tracking, and 3D visualization to recognize parts in the worker’s environment and present relevant information directly through an augmented reality interface. Depending on the available production data, the displayed information may include part category, quantity and customer-order associated with the detected parts
• The system will support workers in locating and counting parts, tracking customer orders, and collecting the information required.. The project will investigate different hardware and algorithmic options for robust detection and identification under realistic shop-floor conditions, including changes in viewpoint, lighting, occlusion, part arrangement, and background
• The final outcome will be a validated proof-of-concept demonstrator that integrates visual perception, part tracking, data retrieval, and an intuitive augmented reality visualization
• Define the system requirements and representative industrial use cases together with the project team
• Test a variety of different AR hardware and libraries
• Set up and document an experimental environment with representative industrial parts and shop-floor conditions
• Acquire and annotate image or video data for part detection, tracking, and identification
• Develop and evaluate computer-vision methods for detecting industrial parts
• Design an augmented reality interface that visualizes part information clearly
• Assess robustness under changes in viewpoint, lighting, occlusion, part arrangement, and background