Particle filter based target tracking from X-bandudnautical radar images

摘要

In this thesis, two particle filter (PF) based visual tracking approaches are designedudfor maneuvering target tracking from X-band nautical radar images: a PF-only basedudapproach and a combined particle-Kalman filters (PF-KF) based approach. Unlikeudexisting Kalman filter (KF) based target tracking algorithms used by nautical radar,udthese two proposed tracking methods both employ a kernel-based histogram model toudrepresent the target in the radar image, and a Bhattacharyya coefficient based similar-udity distance between reference and candidate target models to provide the likelihoodudfunction for the particle filtering. However, the PF-KF method applies a samplingudimportance resampling (SIR) particle filter to obtain preliminary target positions,udand then a Kalman filter to derive refined target positions and velocities. Moreover,udseveral strategies are also proposed to improve the tracking accuracy and stability.udThese strategies include an enhanced reference target model construction method,udupdating reference target model, and adaptive KF for maneuver. Comparison ofudthe target information obtained by the proposed PF-KF method from various fieldudX-band nautical radar image sequences with those measured by GPS shows the pro-udposed approach can provide a reliable and flexible online target tracking for nauticaludradar application. It is also shown that, in the scenario of strong sea clutter, theudproposed approach outperforms the PF-only based approach and the classical track-uding approach which combines order-statistics (OS) CFAR processing and the Kalmanudfilter.