Feature extraction of ship radited noises based on wavelet transform

Wavelet transform was used for extracting features of ship radited noises in this paper.Firstly, Mallat ajgorithm was employed to decompose signals. Secondly, features based on power spectrum and wave shape of the decomposed signals were extracted. Finally, these features were compared with the one of original signal in a neural network classifier. Results show that wavelet transform can enhance the spectrum and shape features of the signal and can reduce the number of dimensions of the feature space.