Target Tracking And Recognition Based On Artificial Fish Swarm Algorithm

The needs of society have led to the continuous development of target tracking and recognition technologies.Target tracking and recognition technologies have been well applied in many fields and have generated tremendous value.Therefore,research on target tracking and identification technology is of great significance.This paper mainly studies the technology of target tracking and recognition.In addition,artificial fish school algorithm is also studied and introduced into target tracking and recognition to improve tracking accuracy and recognition rate.The main research work is as follows:(1)Aiming at the problem that the artificial fish school algorithm adopts a fixed field of view and step length,which causes the oscillation of the algorithm in the later period of convergence and cannot accurately reach the global extremum,two improved artificial fish swarm algorithms are proposed.The improved algorithm adopts a variable horizon and step size.The visual field and the step length are determined according to the distance between the artificial fish and the visual field and the step length are gradually reduced,avoiding the appearance of the artificial fish school algorithm using a fixed small field,small step or large field of view,and large steps.In various cases,experiments have shown that the improved algorithm can converge to the extreme value more accurately.(2)In view of the fact that the LBP feature extraction algorithm is sensitive to noise and the number of binary patterns is excessive,an improved LBP algorithm is proposed in this paper.The number of binary patterns generated by the improved algorithm is reduced by half compared to the original and the noise is increased at the same time.The degree of tolerance,compared with the analysis of the other four feature extraction algorithm,after experiments proved that the improved algorithm has been significantly improved.(3)The particle weight value of the particle filter algorithm will gradually decrease with the iteration and the variance of the particle weight value becomes larger,which leads to the increase of the estimated state and the true state.In this paper,second improved artificial fish swarm algorithms are used in the optimization of particle filtering.According to the optimization principle of the artificial fish school algorithm,the particles are used as artificial fish,and the artificial fish school algorithm can drive the weight of the particles to become larger,and the estimated state is closer to the real state.