Underwater target recognition based on wavelet packet entropy and probabilistic neural network

摘要

A method for underwater target recognition based on wavelet packet entropy and probability neural network is studied in this paper. Wavelet packet transform (WPT) is a time-frequency analysis tool which is developed from wavelet transform (WT). The low-frequency and high-frequency component of a non-stationary signal can be decomposed by WPT simultaneously. The radiated noise of an underwater target is decomposed by WPT and the entropy of terminal nodes through WPT decomposition was selected as feature vector, and is input into a probability neural network (PNN) for underwater target recognition. Simulation result indicates that selecting the entropy as feature vector has higher recognition accurate ratio.