Application of Wavelet Transformation and Neural Network to Magnetic Flux Leakage Signal Classification

According to the non-stationary characteristics of pulse echo signals of flaw in magnetic flux leakage testing system, a method of flaw classification based on the wavelet transform and radius basis function(RBF) neural network was presented. An experiment system was designed to test the method, at first, the feature of flaw was extracted with wavelet transform, then the signal features were classified with RBF neural network, and the fuzzy clustering algorithm was used to find the center of basis function. Experiments showed that the result of recognition was satisfactory and high accuracy of flaw classification could be obtained.