Abstract: The reconstruction of crack profiles is very important in the NDE（nondestructive evaluation） of critical structures,such as pressure vessel and tubes in heat exchanges.First a wavelet transform signal processing technique is used to reduce noise and other nondefect signals from the signals of crack,and then based on an artificial neural network method,the crack profiles are reconstructed.Although the results reveal that this method is with many advantages such as a short CPU time and precision for reconstruction,it does have some drawbacks,for example,the database generation and network training is a much time consuming work.Moreover,this approach does not expressly reconstruct the distribution of conductivity inside a crack,so the reliability of a reconstructed crack shape is unknown.But in practical application,if we do not consider the multiple cracks,this method can be used to reconstruct the natural crack.