Visual Tracking Based On Swarm Intelligent Optimization Algorithm

With the rapid development of science and technology,people hope that computers will be more and more intelligent and can imitate human capabilities.The computer vision discipline is dedicated to making computers have the same visual capabilities as humans.The improvement of computer processing capabilities and the continuous updating of various types of video image acquisition equipment have enabled computer vision technology to have a broader space for development.Object tracking,as one of the most important branches in the computer vision discipline,is widely used in various fields such as video surveillance,human-computer interaction,medical image processing,intelligent transportation,and vision-based control.The process of object tracking can be described as follows: the initial state of a specified object in a certain frame of a given video(the state includes position and size,etc.),and the purpose of tracking is to analyze and evaluate the state of the target object in the next frame of the video.Object tracking can be defined as a priori template information of a known object,and the process of continuously acquiring the object motion state information(such as position,velocity,size,etc.)in the video sequence.Therefore,the object tracking in each frame of video images can be modeled as a problem of optimization.The swarm intelligence optimization algorithm has achieved considerable development over the past few decades.The optimization algorithms represented by particle swarm optimization(PSO),differential evolution algorithm(DE)and artificial bee colony algorithm(ABC)have been widely applied in many fields because of their advantages such as simple structure,few parameters and strong search ability.The main task of this paper is to apply the swarm intelligence optimization algorithm to object tracking to achieve a satisfactory tracking effect.The main work is as follows:Firstly,this article begins with the introduction of the principles of traditional swarm intelligence optimization algorithms and summarizes the advantages and disadvantages of each algorithm.At the same time,the object tracking problem of the video sequence is introduced in detail,as well as an excellent optimization modeling,the structural similarity index(SSIM),which will be used as the fitness function of the optimization algorithm.According to the characteristics of the model,the appropriate swarm intelligent optimization algorithm is selected for further improvement.Secondly,a visual tracking framework based on DE algorithm is proposed.Concerning the fast convergence rate and strong global search ability of DE,the iterative formula and initialization method are modified to make it more conform to the object tracking problem.A new update mechanism is proposed,which can effectively compensate for the errors in the tracking process.Through comparison with other traditional algorithms,it is proved that the framework can achieve a satisfactory tracking effect.For the ABC algorithm's strong global search ability and slow convergence speed,the original iterative formula was modified to speed up the algorithm's convergence speed and enhance the local search ability.This article proposed a brand-new mechanism for the scout bee phase,which enhanced the global search ability of the scout bee so that the overall algorithm improves the convergence speed while the original features of strong global search capability are preserved and improved.Finally,the improved algorithm is combined with the object tracking problem to obtain satisfactory results.This paper mainly based on intelligent optimization algorithm to solve the object tracking problem of the video sequence.Experiments show that the application of swarm intelligence optimization algorithm in the target tracking problem is feasible,and the effect is satisfactory.