Tracking and plot-to-track association for surveillance of busy inland river traffic

River Surveillance systems employ primarily radar sensors, which completely cover an area of interest. Based on the raw videos provided by these radars, plot extraction, plot-to-track association and tracking are then carried out. However, the task of successfully supervising congested traffic areas needs to address ambiguous cases due to vessel maneuvers, radar shadow zones, vessel crossings, etc. This paper deals with the relevance of filtering and data association techniques to resolve these confusing cases.