Abstract
The remote passive detection of vessels in the oceans is a significant activity for improving port security and the security of coastal and offshore operations. There still needs to be an efficient approach to achieve weak ship signal detection with nonparametric and noninformation priors. This study proposes a new multiscale correlation network construction method to effectively distinguish the ship from the ambient noise, which should be promising. Meanwhile, to effectively characterize the constructed network, we render definite the topological network matrix positive definite, then introduce the matrix into the Riemann space to measure the distance between the topology matrix of the noise and the signal by using the geodesic distance. Those methods are demonstrated by simulation and applied to actual recorded data. Compared with the existing network construction and characterization methods, the results show that multiscale correlation network and geodesic distance (GD) methods can distinguish nonlinear time series from noise more effectively.
Original language | English |
---|---|
Pages (from-to) | 992-1008 |
Number of pages | 17 |
Journal | IEEE Journal of Oceanic Engineering |
Volume | 49 |
Issue number | 3 |
DOIs | |
State | Published - 2024 |
Keywords
- Geodesic distance (GD)
- marine environment
- multiscale correlation network (MCN)
- passive ship detection
- remote sonar