Neuro-Fuzzy Modeling Applied to Can Seaming In-Line Inspection

P. Mariño, C.A. Sigüenza, V. Pastoriza, M. Santamaría, E. Martínez, and F. Machado (Spain)

Keywords

In-line Inspection, Statistical Quality Control, Neuro-Fuzzy Modeling, Computer Vision.

Abstract

The authors have been involved in developing an inspection system, based on machine vision, to detect seaming defects in metal cans for fishing food. This paper presents a model building, from the information obtained by the vision system, to take the acceptance/rejection decision for each can. Can seaming process, conventional method for quality inspection of can seaming, and the main dimensional features, are firstly described. Then, the modeling, that includes to generate representative input-output data sets, and mathematical modeling method selection is discussed. Finally, the built model is shown, and its performance is compared with the results of the conventional method.

Important Links:



Go Back