Automatic target tracking method and system by combining multi-characteristic matching and particle filtering

Abstract

The invention relates to an automatic target tracking method and system by combining multi-characteristic matching and particle filtering. The target tracking system provided by the invention comprises a video acquisition module, a tracking algorithm computation module and an output control module, wherein the video acquisition module finishes the initialization of an acquisition card and the real-time acquisition of images; the tracking algorithm computation module comprises three tracking modes: the particle filtering tracking based on gray template region matching, the particle filtering tracking based on color probability distribution and the particle filtering tracking based on SIFT (scale invariant feature transform) characteristic matching, thereby realizing the target tracking of translation space and affine space; and the output control module utilizes the center of the target position obtained by tracking as a control command which is transmitted to a cloud deck, thereby realizing the motion of a camera along with a target object.

Description

Technical field [0001] The present invention relates to a target tracking method and system combining multiple feature matching and particle filtering, in particular to realizing target tracking in translation space and affine space, including particle filter tracking based on gray-scale template region matching and particle filtering based on color probability distribution There are three methods of tracking and particle filter tracking based on SIFT feature matching. Background technique [0002] Target tracking results contain a large amount of temporal and spatial information of moving individuals in the scene, in many aspects such as military visual guidance, robot visual navigation, safety monitoring, traffic control, medical diagnosis, virtual reality and battlefield alerting, public safety monitoring, video compression, and weather analysis. Both have a wide range of applications. [0003] Since the 1980s, scholars at home and abroad have proposed many algorithms for targe...

AI Technical Summary

Problems solved by technology

But its disadvantage is that the accuracy of motion analysis depends on the accuracy of the geometric model, and it is very difficult to obtain accurate geometric models of all moving objects in real life

This limits the use of model-based tracking algorithms. At the same time, tracking algorithms based on 3D models often require a lot of computing time, and it is difficult to achieve real-time moving target tracking.

Images