Computer Aided Quality Control

Quality control in industrial production is a key factor for the success of the products. In order to achieve uniform quality standards, automisation using machine vision methods is becoming more and more important. Often real-time requirements create algorithmic challenges.

• Defect Detection in Wood Surfaces
  In furniture production it is important to use veneer of high quality at all visible surfaces. We have developed anisotropic diffusion filters that allow to identify defects by a simple thresholding step [1], [2].
  
  
  
  
• Visualisation of Dominant Stripes in Nonwoven Farbics
  Nonwoven fabrics appear in numerous applications, ranging from baby napkins over clothing industry to the agricultural sector. One quality parameter is given by the anisotropy caused by dominant stripe-like structures. Such structures can be made visible by suitably adapted anistropic diffusion processes [1], [2].
• Multiscale Analysis of the Cloudiness of Nonwoven Fabrics
  The cloudiness constitutes another crucial quality parameter for nonwoven fabrics. It is a scale phenomenon. Therefore, we have developed a tool for automatic quality assessment taking into account the cloudiness at multiple scales. It involves specific pyramid decomposition and allows real-time assessment of an entire production line with a simple PC [3], [4].
  This algorithm has entered the commercial sofware product MASC-VQC of the Fraunhofer Institute for Techno- and Economathematics.
• Multiple View Inspection of Aluminium Die Castings
  While classical automated multiple view inspection requires calibrated cameras, we propose a novel method that can also cope with uncalibrated images. It tracks defects of industrial objects along an image sequence. The success of this strategy is demonstated by applying it to the inspection of aluminium die castings [5].
• Automated Defect Detection in Wine Bottlenecks
  For the inspection of wine bottlenecks, a novel system is constructed that uses an inner lighting source. Computer vision methods are applied for tracking potential defects along an image sequence. The current implementation of the system detects 87 percent true positives and 0 percent false positives [6].

1. J. Weickert:
   Anisotropic diffusion filters for image processing based quality control.
   A. Fasano, M. Primicerio (Eds.): Proc. Seventh European Conf. on Mathematics in Industry. Teubner, Stuttgart, 355-362, 1994.

2. H. Neunzert, B. Claus, K. Rjasanowa, R. Rösch, J. Weickert,
   Mathematische Werkzeuge in der Bildverarbeitung zur Qualitätsbestimmung von Oberflächen.
   K.-H. Hoffmann, W. Jäger, T. Lohmann, H. Schunck (Eds.), Mathematik - Schlüsseltechnologie für die Zukunft. Springer, Berlin, 449-462, 1997.
   

3. J. Weickert:
   A model for the cloudiness of fabrics:
   H. Neunzert (Ed.): Progress in industrial mathematics at ECMI 94. Wiley-Teubner, Chichester, 258-265, 1996.

4. J. Weickert,
   A real-time algorithm for assessing inhomogeneities in fabrics,
   Real-Time Imaging, Vol. 5, 15-22, 1999.

5. L. Pizarro, D. Mery, R. Delpiano, M. Carrasco:
   Robust automated multiple view inspection
   Pattern Analysis and Applications, Vol. 11, No. 1, 21-32, 2008.
   Revised version of Technical Report No. 192, Department of Mathematics, Saarland University, Saarbrücken, Germany, April 2007.

6. M. Carrasco, L. Pizarro, D. Mery:
   Image acquisition and automated inspection of wine bottlenecks by tracking in multiple views.
   Proc. of the 8th International Conference on Signal Processing, Computational Geometry and Artificial Vision - ISCGAV 2008, pp. 82-89, August 2008.

MIA Group
©2001-2023
The author is not
responsible for
the content of
external pages.

Imprint - Data protection