Research On Infrared Multi-Target Tracking Algorithm Based On Particle Filter

In recent years,with the rapid development of information technology,especially the rapid development of computing technology and infra-red Imaging technology,infrared target detection and tracking technology in military and civilian plays a more important role.Define multiple target tracking technology as found in real time multiple targets,and target status parameter information,based on the target information analysis and estimation of intention and situation.Especially in tracking multiple targets,multiple objectives covered and track real-time performance and other issues become the focus of research is also the difficulty,there are many problems need to be addressed or improved.Over the years,particle filter algorithm for nonlinear filter is the most important kind of filtering methods,have achieved great progress.First,this paper analyzes the classical particle filter algorithm theory and classical particle filtering algorithm and its application in infrared target tracking and multiple target tracking related problems of mathematical modeling.Next in a Bayesian tracking theory discusses the advantages and disadvantages of the structure.Secondly,it describes the simple theory of mean shift algorithm,and then presents a mean shift algorithm combined with particle filter tracking algorithm for infrared object tracking method to use.This method maintains the calculation of mean shift algorithm,and has the characteristics of a good real-time.The mean shift algorithm to converge the particles in a particle filter so that each particle has a more realistic target,greatly reduces the number of particles needed to describe the target State,and enhance the efficiency of particle sampling,improved algorithms in real-time.Through several experiments show that track fusion algorithm with robust performance,and save time taking into account the demands of real-time target tracking.Finally,research on improved particle filter multi-target detection algorithm.Describes a common algorithm for infrared target detection;analysis of advantages and disadvantages on the basis of its own,a new optimization algorithm for infrared target detection based on weight,improved resampling methods,making weight large particles enter the next track instead of weights smaller particles,improve the accuracy of tracking and persistence.Subsequently,introduced in weights optimized particle filter algorithm based on Markov random,undirected graph model to handle multiple targets tracking data Association problem,makes the algorithm of blocking multiple targetenhance the effectiveness of tracking of the target when it is partially occluded.This paper discusses the improved particle filter algorithm and its real-time tracking and anti-blocking than common particle filter has improved greatly,multiple target tracking technology has some theoretical significance and applied value.