cross-evaluation of a parallel operating svm – cnn classifier for reliable internal decision-making processes in composite inspection参考文献.pdf

Journal of Manufacturing Systems 60 (2021) 620–639
Contents lists available at ScienceDirect
Journal of Manufacturing Systems
journal homepage:
Technical Paper
Cross-evaluation of a parallel operating SVM – CNN classifier for reliable
internal decision-making processes in composite inspection
Sebastian Meister a,b,*, Mahdieu Wermes a, Jan Stüve a,b, Roger M. Groves b
a Center for Lightweight Production Technology (ZLP), German Aerospace Center (DLR), Ottenbecker Damm 12, Stade 21680, Germany
b Aerospace Non-Destructive Testing Laboratory, Delft University of Technology, Kluyverweg 1, Delft 2629, The Netherlands
ARTICLE INFO ABSTRACT
Keywords: In the aerospace industry, automated fibre laying processes are often applied for economical composite part
Explainable Artificial Intelligence fabrication. Unfortunately, the current mandatory visual quality assurance process takes up to 50% of the entire
Automated Fiber Placement manufacturing time. An automised classificationof manufacturing deviations using Neural Networks potentially
Inline inspection
improves the inspection’s effectiveness. Unfortunately, the automated decision-making procedures of machine
Convolutional Neural Network
learning approaches are challenging to trace. Therefore, we introduce an approach for evaluating the classifiers
Laser Line Scan Sensor
Support Vector Machine response for this use case.
For this purpose, we present a parallel classification approach of Convolutional Neural Network (