An Adaptive Particle Swarm Optimization for Underwater Target Tracking in Forward Looking Sonar Image Sequences

In order to obtain accurate underwater target tracking results in forward-looking sonar image sequences, an adaptive particle swarm optimization (APSO) algorithm is proposed to track the underwater target. Specifically, an adaptive inertia weight is first applied to solve the problem of diversity loss and premature convergence in the PSO algorithm. Then, the individual optimal fitness value of a random particle selected from particle swarm is compared with that of the current particle, and the larger one is used to update the velocity of the particle to further prevent particles from falling into the local optimum. On this basis, when the underwater target is occluded, a new update strategy based on an adaptive discrete swarm optimization algorithm is used to update the particle's position according to the level of occlusion so as to achieve a better tracking result. Finally, the experimental results show that the proposed APSO can accurately complete underwater target tracking. Compared with other algorithms, the proposed APSO has higher tracking accuracy and faster tracking speed, and it has a certain effectiveness and adaptability.