Frequency line tracking on a lofargram: an efficient wedding between probabilistic data association modelling and dynamic programming technique

This paper deals with the problem of frequency line extraction in a lofargram which is encountered in a passive sonar system. Using adequate probabilistic data association (PDA) modelling and the principle of dynamic programming, we propose a new algorithm of frequency line tracking in the presence of false alarms. Results are presented for several signal to noise ratios: the results show that this new algorithm is able to extract a frequency line correctly when the signal to noise ratio is higher than 4 dB on the lofargram (or -23 dB on the temporal signal, before a 512 point FFT).