Loading...

FEM enhanced signal processing approach for pattern recognition in the SQUID based NDE system

Sarreshtedari, F ; Sharif University of Technology

564 Viewed
  1. Type of Document: Article
  2. DOI: 10.1088/1742-6596/234/4/042030
  3. Abstract:
  4. An efficient Non-Destructive Evaluation algorithm has been developed in order to extract the required information for pattern recognition of defects in the conductive samples. Using high-Tc gradiometer RF-SQUIDs in unshielded environments and incorporating an automated two dimensional non-magnetic scanning robot, samples with different intentional defects have been tested. We have used a developed noise cancellation approach for the improvement of the effectiveness of the used inverse-problem technique. In this approach we have used a well examined Finite Element Method (FEM) to apply a noise reduction filtering on the obtained raw magnetic image data before incorporating the signal processing analysis. By applying this noise cancellation filter and incorporating three different signal processing algorithms and comparing the results of the predicted images by the pattern of the intentionally made defects, we have investigated the ability of these methods for pattern recognition of unknown defects
  5. Keywords:
  6. Gradiometers ; High-Tc ; Magnetic images ; Noise cancellation ; Noise cancellation filters ; Noise reductions ; Non destructive evaluation ; Nonmagnetics ; Processing approach ; Signal processing algorithms ; Unshielded environment ; Algorithms ; Defects ; Finite element method ; Signal processing ; Spurious signal noise ; Superconductivity ; Pattern recognition systems
  7. Source: Journal of Physics: Conference Series, 13 September 2009 through 17 September 2009 ; Volume 234, Issue PART 4 , 2010 ; 17426588 (ISSN)
  8. URL: http://iopscience.iop.org/article/10.1088/1742-6596/234/4/042030/meta