Robust group target tracking method and tracking system for monitoring system

Abstract

The invention provides a robust group target tracking method for a monitoring system. The method comprises the following steps: S1, establishing a hybrid observation model; S2, calculating the joint probability density of each state variable and the measured complete data of the monitoring system; S3, calculating the approximate posterior probability density; S4, calculating clutter density parameters in measurement; S5, according to the approximate posterior probability density of each state variable and a clutter density parameter in measurement, calculating the mathematical expectation of each state variable so as to obtain an estimated value of each state variable, repeatedly executing the steps S2 to S4 for appointed times, and considering that tracking is finished; and S6, performingtracking estimation of the group target: realizing estimation of the motion state, the shape and the environment clutter density of the group target according to the finally obtained estimation valueof each state variable and the clutter density parameter. According to the method, the problems caused by observation noise obeying heavy tail distribution and environmental clutters are solved, andmeanwhile, the clutter density is estimated when the group target is tracked.

Description

technical field [0001] The present invention relates to the technical field of robust group target tracking methods, in particular to a monitoring system-oriented robust group target tracking method, especially a random matrix model-based robust group target tracking method. Background technique [0002] The definition of a group target is: under the condition of satisfying a certain distance between targets, the spatial relationship (such as position, etc.) is relatively fixed (or has similar state parameters) for a certain long period of time, and there are at least two distinguishable or indistinguishable targets Target clusters formed by mixing. Swarm object tracking needs to simultaneously solve the problem of joint estimation of the overall motion and shape of the swarm. A swarm target tracking method based on a random matrix model uses a symmetric positive definite matrix to characterize the shape of the swarm. That is, on a two-dimensional plane, the shape of the g...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects in the prior art, the purpose of the present invention is to provide a monitoring system-oriented robust group target tracking method and tracking system to solve the problems caused by observation noise and environmental clutter subject to heavy-tailed distribution, and at the same time The clutter density is estimated when the group target is tracked, and the joint estimation of the shape, motion state and environment of the group target is realized when the sensor observation noise has a heavy-tailed distribution characteristic and produces abnormal measurements, and there is clutter interference in the monitoring environment Clutter density

Images