Abstract: Nondestructive evaluation of ferromagnetic material is most commonly performed by magnetic flux leakage (MFL) techniques, and it is key element to describe the characters of defects from MFL detecting signals. A new method for the reconstitution of 2D profiles is presented based on least squares support vector machines (LS-SVM) technique, the input data set of SVM is MFL signals and output data set is 2D profiles parameter, the mapping relationship from MFL signals to 2D profiles of defects is established. The least squares method is introduced into network learning, the training data set is composed of experiment data set and emulational data set, the testing data set is artificial crack defects. The reconstitution of 2D profiles of artificial crack defects in the magnetic flux leakage detecting is implemented by this algorithm. Compared with the reconstitution results of RBF network, the results show that LS-SVM possesses quickens speed, enhances high accuracy and is of very good generalization ability , and it is a good way for the quantization of the MFL detecting.

Key words: magnetic flux leakage detecting, LS-SVM, 2D profiles, defect, reconstruction, pipeline