Extraction of time-frequency target features
- Resource Type
- Conference
- Authors
- Oesterlein, Tobias G.; He, Chensong; Quijano, Jorge E.; Campbell, Richard L.; Zurk, Lisa M.; Siderius, Martin
- Source
- 2010 Conference Record of the Forty Fourth Asilomar Conference on Signals, Systems and Computers Signals, Systems and Computers (ASILOMAR), 2010 Conference Record of the Forty Fourth Asilomar Conference on. :2156-2163 Nov, 2010
- Subject
- Communication, Networking and Broadcast Technologies
Signal Processing and Analysis
Computing and Processing
Target tracking
Spectrogram
Kalman filters
Sonar
Noise
Time frequency analysis
Equations
- Language
- ISSN
- 1058-6393
Physics-based detection algorithms can improve discrimination of sonar targets from competing bottom reverberation, but are vulnerable to environmental uncertainties. Recent research in the underwater community has identified an environmentally robust time-frequency signature for improved target discrimination. Application of this “invariant” requires processing algorithms to identify striations in a spectrogram and to quantify the associated track certainty. In this paper, two robust invariant-based algorithms are presented and demonstrated with underwater data. The first algorithm uses a Kalman Filter to estimate the time-frequency striations in sonar spectrograms. The second computes a “likeliness” metric to measure discrimination between target and non-target detections.