A Classification Method of Magnetic Flux Leakage Signal for Internal and External Wall Defects in Seamless Steel Pipes Based on Spectrogram and Support Vector Machine

Magnetic flux leakage non-destructive testing is an important method for comprehensive inspection of seamless steel pipes. Distinguishing the signals formed by internal and external wall defects is a significant issue in this method. This study proposes a new approach, which first converts the magnetic flux leakage signal obtained by the sensors into a spectrogram and calculates its gray level co-occurrence matrix. Then, it calculates the energy, entropy, and correlation at four different angles of the matrix. The mean and standard deviation of these quantities are used as the characteristic values of the defect signal. Combining with support vector machine technology, an identification model for the signals of internal and external wall defects is established. The model processes the test signal to achieve automatic defect classification. Experimental results show that this method can effectively identify and classify defect signals with high accuracy.