New Signal Processing Scheme for the Analysis of Electromagnetic Images

Eddy Current Imaging, Array Sensors and Flaw Reconstruction

“…The compressed data set is treated as a feature vector and conventional pattern recognition techniques can be used for classifying it in the feature space containing feature vectors of defects prototypes. The main disadvantage of the method proposed in [9] is that the method does not rely on any phenomenological model and the dimension of the feature space may be very large if defects of different types are searched for. Therefore, it seems reasonable to combine a singular value decomposition method with other flaw reconstruction techniques based on models of the probe response to a flaw.…”

“…is known, i.e., it is calculated taking into account the (k -I)-th iteration of the perturbed electric field. The k-th iteration of the perturbed electric field is a solution of the following set of partial differential equations: in conducting semi-space (z < 0), and (8) in nonconducting semi-space ( z > 0), with the boundary conditions (9) and at the plane z = 0. The polarization current density Fk) is calculated from:…”

“…Another approach to the eddy current image restoration and defect characterization has been presented in [9]. The philosophy of the method described there is somewhat similar to the Fourier descriptor approach for the classification of eddy current impedance plane trajectories [2-41. The eigenvalues and corresponding left and right eigenvectors of the image matrix are computed by means of standard singular value decomposition procedures and the image is resynthesized from first several singular values and corresponding singular vectors.…”