Problems of signal recognition in ultrasonic inspection are analyzed. A new fusion recognition method base on multiple features extraction is researched, which compounds with support vector machine theory and Bayes reasoning. The principles of SVM method and the Bayes reasoning are introduced. Fusion recognition method base on maximum a posteriori(MAP) are designed to identify the signals of different flaws with features extracted from different ways. Four feature extraction methods from different spatial domains of a signal are presented for fusion recognition. Experiments with both SVM method and SVM-Bayes method are carried out to identify the flaw signals of oil casing pipe. The result shows that flaws can be identified effectively by SVM-Bayes method, and both recognition correct rate and generalization are better than a single feature SVM method.
Che Hong-kun;Lu Fu-zai;Xiang Zhan-qin.
Ultrasonic signal recognition by multiple features svm-bayes fusion method[J]. Journal of Vibration and Shock, 2011, 30(12): 265-269