Dense multiple-target tracking based on dual frequency identification sonar (DIDSON) image

This paper used DIDSON to get information of underwater fish and carried out the real-time multiple-target tracking, particularly the targets were dense. Analysis of fishery resources was also done through the data post-processing. Realtime data processing mainly included DIDSON image building/enhancement/filter, target detection and multiple-target tracking. Nearest neighbor clustering algorithm combined with Extended Kalman Filter was applied to the multiple-target tracking. Data post-processing mainly implemented the statistics of fish duration, swimming speed, swimming direction, trajectory length, which will provide strong technical support for the fishery resources assessment.