Defect identification based on wavelet decomposition for MFL non-destructive inspection of steel plates

The article considers an application of wavelet decomposition for defect identification based on the example of magnetic flux leakage (MFL) non-destructive testing (NDT). To solve this problem, a parallel signal decomposition is performed for several chosen types of mother wavelet. This article describes the criteria and algorithm for the choice of mother wavelet. A decision tree is used for defect classification. The algorithm presented has been tested both on imitations of defects and on an oil tank floor plate with natural defects.