Study on Ultrasonic Flaw Identification Based on Complex Wavelet Transform and Bayesian Clssifier
Shen Xiao-an · Modular Machine Tool & Automatic Manufacturing Technique · 2008
According to the classic indeterminate problem of flaw detection,a new approach of flaw identification was presented in this paper based on complex wavelet transform and Bayesian classifier.On the basis of de-noising of the defect echoes by utilization of wavelet theory,the envelope of defect echo signal,which was then used to extract the characteristic parameters,was obtained based on complex wavelet transform..Finally,the Bayesian classifier was used.The experiment results indicate that the presented technique is very useful in the field of flaw detection.