Particle filter integrated tracking method based on support vector machine

Abstract

The invention discloses a particle filter integrated tracking method based on a support vector machine. The particle filter integrated tracking method based on the support vector machine comprises the following steps: establishing a support vector machine judging model based on a global gray feature; classifying candidate targets obtained by particle sampling; outputting candidate particles with higher reliability and taking the candidate particles with higher reliability as input of a block sparse generating model; establishing a generating model based on block sparse representation; shielding the input candidate targets and carrying out similarity measurement on the input candidate targets on the basis of reference information; and meanwhile, keeping apparent change of the targets updated by updating training samples of the support vector machine and the reference information of the block sparse model so as to realize stable tracking.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a video target tracking method, which can be used to realize precise tracking of targets under different interference factors. Background technique [0002] Automatic target tracking based on image and video sequences is an important content in the field of machine vision and pattern recognition. It has been widely used in intelligent monitoring, visual navigation, video retrieval and other fields. At present, a lot of research has been done on the robust tracking of targets at home and abroad. However, in practical applications, it is still a very challenging task to design an accurate, stable, and real-time video tracking algorithm due to factors such as target appearance changes, attitude changes, imaging environment changes, background interference, and occlusion. [0003] Target tracking algorithms are mainly divided into two categories: discriminative and generati...

AI Technical Summary

Problems solved by technology

However, this method still has the following shortcomings: First, the contributions of the discriminative model and the generative model are not balanced enough in the tracking process. In most cases, only the generative model can achieve the effect close to that of the joint model.

That is to say, the joint model does not make full use of the separable information of the discriminant model; secondly, the discriminant model and the generative model in this method process a large number of candidate particles separately in terms of logical structure, and finally synthesize their respective results. The amount of calculation is too large, which affects the real-time effect

Images