Multiple Target Tracking Method Research Based On The Static Background

With the rapid development of the field of computer vision, Target Tracking Method Based on the static background in computer science nowadays has become a hot research.This paper studies research under multi-target tracking method to video images of a static background for the study, followed by simulation of the image simulation. Target tracking technology in the visual navigation, intelligent control, human-computer interaction and precision-guided and other fields have a wide range of applications. Study on static background image target tracking method has its significance and practical application value.VIBE detection method has recently been proposed as a target detection method based on background model, with its intuitive and fast features are subject to the concern of many scholars at home and abroad, to VIBE detection algorithm in-depth study on the development of target detection technology is there significance. Multi-target tracking area is the focus of this study, multi-target tracking in real life has broad application prospects, but their research is more complex,this paper commonly used to study particle filter,on the basis of its sequential Monte Carlo Luo PHD algorithm study and discussion.This paper describes the status and significance of multi-target tracking under static background, followed by the introduction of the common target detection method, described in detail Vibe to the algorithm, and its improvements. Particle filter has a relatively other algorithms can be applied to non-linear, non-Gaussian capacity of the system, we use this method to achieve a video single, multi-target tracking problem. The particle filter for particle filter method is applied with a multi-target tracking can only track the number of known targets, time-consuming computational complexity associated with data limitations, this article introduces the PHD algorithm that can handle an unknown number of targets, the algorithm and PHD It does not require data association process, so the speed of this method is better than the multi-target tracking process based on particle filtering and data association methods,and will apply it to an analog image simulation. This paper introduces the advantages of Science magazine a concise excellent clustering method -Clustering by fast search and find of density peaks, the algorithm has no input parameters compared to those of K-means clustering method, etc., and its simple, quick .