By Paolo Franchi
Big data exploitation for active sonar contact-level classification. De Magistris, Giovanni; Stinco, Pietro; Canepa, Gaetano; Ferri, Gabriele; Fois, Marco; Tesei, Alessandra; LePage, Kevin D. CMRE-FR-2019-005. May 2020.
Neural networks are proposed in this report to classify mobile targets from active sonar sensor data for underwater surveillance. The raw signal is processed, transformed in the time-frequency domain and classified (target/no target). The values of the neural network parameters (weights and biases) are learned using data collected during two sea trials with an Echo-Repeater as a surrogate target. The classifier is then validated using data from a third sea trial in different geographical locations and environmental conditions. In the validation dataset used in this report, the CNN classifier significantly reduces the number of false alarms and outperforms the traditional feature-based classifier that was previously developed.