Underwater echo signal wake envelope feature extraction based on discrete wavelet transform

A feature extraction method for underwater target echoes is proposed. Based on the discrete wavelet transform, this method can classify the underwater substrate based on the scattering components in the echo signal, extract the underwater echo signal wake envelope feature as the identification feature vector, constitute the feature space based on the wake envelope feature, and then use the best identification vector method to compress the feature space, and finally use the minimum distance classifier to classify the target echo. The measured data show that the envelope feature obtained by this method is a robust and effective feature, which can obtain a high correct recognition rate.