Automatic recognition of underwater acoustic signature for naval applications
Résumé
In a military context, where human capabilities are no longer sufficient to process quickly and reliably an ever-increasing amount of data, the implementation of algorithms based on Artificial Intelligence (AI), through the computing power of modern infrastructure, increases the ability to interpret and correlate massive heterogeneous data. This article will present an original automatic underwater acoustic signature recognition technique. The experiments are carried out from public underwater acoustic dataset. Besides, the performance of three architectures based on Mel-frequency cepstral coefficients (MFCCs), Machine Learning techniques and neural networks are compared.