Underwater acoustic target recognition based on Gammatone filterbank and instantaneous frequency

The performance of underwater acoustic target recognition system decreases significantly mainly due to the complex underwater acoustic conditions. Nevertheless, human auditory system is capable of segregating different targets in noisy conditions. Several auditory-based features have been applied to underwater acoustic target recognition, but only represent amplitude of the signal. However, phase represented by the instantaneous frequency (IF) might also contain some information to distinguish different targets. In this paper, a novel feature extraction method based on the Gammatone filterbank and instantaneous frequency is applied to robust underwater acoustic target recognition. The Support Vector Machine (SVM) is employed as the classifier in experiments. Experimental results indicate that the novel feature extraction method has a robust performance than conventional methods for underwater acoustic target recognition.