A LBF-associated contour tracking method for underwater targets tracking

To improve the tracking robustness in fuzziness when the underwater targets are overlapped in forward-looking sonar images, a contour tracking method that use Local Binary Fitting (LBF) to constrain the evolution of particle filter is discussed for underwater targets tracking in this paper. A space prior is generated from underwater target contour information which is used to restrict the boundary particles in the resampling process of particle filter. The relation between the active contour and moving targets is established, and the space prior is defined according to the distance between particles and the detected contour. Experimental results show that LBF-associated contour tracking algorithm has better expression on target contour in comparison with conventional tracking methods.