Detection of damaged beverage crates by means of neuronumerics

摘要

The inspection of returned crates of beverages as well as bottles is mainly performed by optical systems in industrial automatic filling lines. Here a powerful novel system for the detection of small and hidden damages is being developed, which is based on the principle of mechanical excitation. The selection of individual crates occurs automatically by an artificial neural network (ANN), which is trained with data obtained both from experiment and from numerics. Thus, using neuronumerics cracks are detected with an accuracy of over 99 %.