Track Before Detect Algorithm for Tracking Extended Targets applied to Real-World Data of X-band Marine Radar

This communication presents the application of a Track before Detect technique for tracking extended targets in cluttered environment using a high resolution marine radar. The algorithm makes use of a particle filtering approach and assumes that the target shape is an ellipse. We compute numerically the posterior probability of the target presence jointly with the estimation of the target state, augmented with its geometrical parameters (length and width), that could be used for target classification. The results of tests with real-world data are reported, and a comparison, when possible, with the ground truth is also provided.