Research on MFL signal WPES feature extraction method for pipeline cracks defection

Magnetic flux leakage (MFL) detection method is an important technology for pipeline defect detection, recent studies show that the defect MFL detection signal in the high frequency signal characteristics is not obvious, the spectrum sequence is very fragmented, high frequency feature extraction difficult, in post-processing, signal distortion and distortion phenomenon occurs. This paper established an improved approach to integration wavelet packet transform and Wigner-Ville transform (WVT) algorithm, using continuous wavelet transform and discrete wavelet transform technology to handle basic mathematical tools and complete the signal characteristics extraction, to achieve fusion analysis of high frequency WVT processing to eliminate the signal post miscellaneous. By this method, wavelet packet energy spectrum as defect MFL detection signal characteristic values, the recognition and extraction conversion eigenvalues for the corresponding signal feature identification and extraction of energy on the realization of the full frequency characteristics of the signal, which will help eliminate signal distortion problems and improve the accuracy of detection signal characteristics analysis and defect identification.