论文信息 - Marine vessels classification using probabilistic neural network

Marine vessels classification using probabilistic neural network

The establishment of intelligent systems for classifying marine vessels due to the acoustic radiated noise is of major importance to the defense systems of many countries. This paper presents a new method for classification of marine vessels which has been examined by a great amount of real data. In this method, the acoustic noise of ships is modeled by AR (Autoregressive) model with suitable order and coefficients of this model are used for classification of ships. The presented algorithm consists of a Probabilistic Neural Nehvork (PNN) where the AR model coefficients are used as the input vector. The evaluation results of proposed method with real data show this method is successful in classifying ships to !\V 0 separate categories (heavy and medium ships).

H. Keivani | H. Rastegar | Mehdi Farrokhrooz

[1] Carl G. Looney,et al. Pattern recognition using neural networks: theory and algorithms for engineers and scientists , 1997 .

[2] J. G. Lourens. Classification of ships using underwater radiated noise , 1988 .

[3] James E. Barger,et al. Underwater acoustic system analysis , 1985, Proceedings of the IEEE.

[4] Robert J. Urick,et al. Principles of underwater sound , 1975 .

[5] David M. Drumheller,et al. Sonar Signal Processing , 1999 .

[6] Jinlin Wang,et al. An application of expert system in recognition of radiated noise of underwater target , 1995, 'Challenges of Our Changing Global Environment'. Conference Proceedings. OCEANS '95 MTS/IEEE.

[7] H. Amindavar,et al. Determining class of underwater vehicles in passive sonar using hidden Markov model with Hausdorff similarity measure , 1998, Proceedings of 1998 International Symposium on Underwater Technology.