Joint estimation of target number and DOA using reversible jump MCMC in sea clutter

We address the problem of detection of a small target in high frequency ocean surveillance radar using the method of reversible jump Monte Carlo Markov chain. The number and the directions of arrival of targets are estimated simultaneously. However, the existence of sea clutter makes the traditional RJMCMC inefficient, and an adaptive preprocessing is preferred in combination with the original RJMCMC to achieve better detection. The performance of the proposed method is analyzed through several experiments with real data.