Fuzzy Inference Systems for Invariant Pattern Recognition in MFL NDE

  • S. Mandayam
  • L. Udpa
  • S. S. Udpa
  • W. Lord


Defect related information present in NDE signals is frequently obscured by the presence of operational variables inherent in the system. A typical NDE system comprises of an energy source, a test specimen and a sensor array. Operational variables include uncontrollable changes in source signal strength and/or frequency, variations in the sensitivity of the sensor and alterations in the material properties of the test specimen. These operational variables can confuse subsequent signal interpretation schemes, such as those relying on artificial neural networks. Invariant pattern recognition methods are required to ensure accurate signal characterization in terms of the underlying defect geometry. This paper describes a generalized invariance transformation technique to compensate for operational variables in NDE systems. An application to magnetic flux leakage (MFL) inspection of gas transmission pipelines is presented. The technique is employed to compensate for variations in magnetization characteristics in the pipe wall.


Membership Function Radial Basis Function Fuzzy Inference System Fuzzy Rule Base Gaussian Membership Function 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    E. Barnard and D. Casasent, “Invariance and neural nets,” IEEE Transactions on Neu ral Networks, Vol 2., No. 5, pp. 498–508, September 1991.CrossRefGoogle Scholar
  2. 2.
    S. S. Udpa and W. Lord, “A Fourier descriptor classification scheme for differential probe signals,” Materials Evaluation, 42(9), pp. 1136–1141, 1984.Google Scholar
  3. 3.
    M. K. Hu, “Visual pattern recognition by moment invariants,” IRE Transactions on Information Theory, Vol. IT-8, pp. 179–187, 1962.Google Scholar
  4. 4.
    N. E. Cotter, “The Stone-Weierstrass theorem and its application to neural networks,” IEEE Transactions on Neural Networks, Vol. 1, No. 4, pp. 290–295, December 1990.MathSciNetCrossRefGoogle Scholar
  5. 5.
    R. С Williamson and U. Helmke, “Existence and uniqueness results for neural net work approximations,” IEEE Transactions on Neural Networks, Vol 6., No. 1, pp. 2–13, January 1995.MathSciNetCrossRefGoogle Scholar
  6. 6.
    L. Wang and J. M. Mendel, “Fuzzy basis functions, universal approximation, and orthogonal least squares learning,” IEEE Transactions on Neural Networks, Vol 3., No. 5, pp. 807–814, September 1992.CrossRefGoogle Scholar
  7. 7.
    B. Kosko and S. Isaka, “Fuzzy logic,” Scientific American, pp. 76–81, July 1993.Google Scholar
  8. 8.
    J. S. Jang and С. Т. Sun, “Functional equivalence between radial basis function net works and fuzzy inference systems,” IEEE Transactions on Neural Networks, Vol 4., No. 1,pp. 156–159, January 1993.CrossRefGoogle Scholar
  9. 9.
    J. A. Dickerson and B. Kosko, “Fuzzy function learning with covariance ellipsoids,” IEEE International Conference on Neural Networks, pp. 1161–1167, March 1993.Google Scholar
  10. 10.
    J. Nie and D. A. Linkens, “Learning control using fuzzified self-organizing radial basis function network,” IEEE Transactions on Fuzzy Systems, Vol. 1, No. 4, pp. 280–287, November, 1993.Google Scholar
  11. 11.
    H. Nomura, I. Hayashi and N. Wakami, “A learning method for fuzzy inference rules by descent method,” IEEE International Conference on Fuzzy Systems, pp. 203–210, March 1992.Google Scholar
  12. 12.
    J. T. Tou and R. С Gonzalez, Pattern Recognition Principles, pp. 94–97, Addison-Wesley, Reading, MA, 1974.MATHGoogle Scholar
  13. 13.
    G. J. Posakony and V. L. Hill, Assuring the Integrity of Natural Gas Transmission Pipelines, Topical Report, GRI 91–0366, Gas Research Institute, Chicago, IL, November 1992.Google Scholar

Copyright information

© Plenum Press, New York 1996

Authors and Affiliations

  • S. Mandayam
    • 1
  • L. Udpa
    • 1
  • S. S. Udpa
    • 1
  • W. Lord
    • 1
  1. 1.Department of Electrical and Computer EngineeringIowa State UniversityAmesUSA

Personalised recommendations