Logo des Repositoriums

Auflistung nach Schlagwort "Quality Control"

1 - 2 von 2
  • Konferenzbeitrag
    Increasing Reliability in FDM Manufacturing
    (INFORMATIK 2019: 50 Jahre Gesellschaft für Informatik – Informatik für Gesellschaft (Workshop-Beiträge), 2019) Heider, Michael
    Additive Manufacturing machines following the Fused Deposition Modelling process can rapidly produce wide varieties of parts. A 3D computer model is divided into instructions the FDM machine uses to produce the part layer by layer. Numerous parameters can be modified to improve the instructions generated and extensive research is being performed into determining optimal parameters. Due to the complexity of the process and limited available data about influence factors, that might change over the duration of manufacturing, some produced parts have subpar quality or fail to be produced at all. An early automated detection that the resulting part will not be inside the preset quality tolerances could save substantial resources by not finishing production on those parts. Furthermore it might be possible to utilise machine learning techniques such as XCS to adaptively change instructions during printing as to return the part into the accepted parameter range.
  • Konferenzbeitrag
    Machine Learning in Glass Bottle Printing Quality Control: A Collaboration with a Medium-Sized Industrial Partner
    (INFORMATIK 2024, 2024) Bundscherer, Maximilian; Schmitt, Thomas H.; Bocklet, Tobias
    In cooperation with a medium-sized industrial partner, we developed and evaluated two ML-based approaches for quality control in glass bottle printing. Our first approach utilized various filters to suppress reflections, image quality metrics for image comparison, and supervised classification models, resulting in an accuracy of 84%. We used the ORB algorithm for image alignment and to estimate print rotations, which may indicate manufacturing anomalies. In our second approach, we fine-tuned pre-trained CNN models, which resulted in an accuracy of 87%. Utilizing Grad-CAM, an Explainable AI method, we localized and visualized frequently defective bottle print regions without explicitly training our models for this use case. These insights can be used to optimize the actual manufacturing process beyond classification. This paper also describes our general approach and the challenges we encountered in practice with data collection during ongoing production, unsupervised preselection, and labeling.