Research On Dynamic Group Target Simulation And Collective Tracking Method Under Space Based Detection Condition

The space-based detection system has the characteristics of wide coverage,fast target discovery,and multiple combat uses.It is the most important link in space defense.The establishment of a complete space-based detection system is the country’s future development direction.The core mission is to track and monitor all-round high-speed flying targets.This subject is aimed at releasing multiple space debris during the flight of a high-speed aircraft,and achieving stable tracking of a group of multiple targets.Firstly,analyze the characteristics of the group target,perform motion modeling and infrared radiation modeling of the group target,and obtain the group target simulation image sequence by analyzing the imaging principle of the sensor;then realize the group target feature based on the local probability distribution for the simulated image sequence Extraction;Finally,the Gaussian mixture PHD filter algorithm is used to achieve the same tracking of group targets.The research content is as follows:(1)Infrared imaging simulation of group targets.Under the conditions of space-based detection,the group target motion model and the group target’s infrared radiation model are established.For the an alysis of the group target motion model,the relative motion method is used to establish the motion equation of the space debris relative to the target body,and the relationship between the space debris and the target body is obtained;for the infrared radiation model of the group target,because the group target flying in space emits outward Thermal radiation absorbs external thermal radiation at the same time.After the target reaches thermal equilibrium,the equilibrium temperature of the high-speed flying group target is calculated.According to Planck’s law,the irradiance of the target reaching the sensor is calculated;for the infrared sensor imaging simulation model,first use the coordinate system To determine the corresponding relationship between the coordinates of the space target and the image plane;use the irradiance information of the target at the camera mouth to determine the focal plane response of the pixel to the target and background,and use STK software to simulate satellite observati on of high-speed flying targets Set the corresponding satellite parameters and sensor parameters,and then realize the dynamic simulation of the group target,obtain the image sequence of the group target release process,and provide data support for the f urther development of the group target tracking.(2)Group target detection and tracking method based on local probability distribution.As the infrared sensor generates noise and clutter,it affects the detection performance of the infrared sensor,especi ally when the target pixel occupies very few,the target will be submerged in the background noise,resulting in missed detection.To solve this problem,consider that in a small area around the target,the target and the neighborhood background will form a sharp contrast.The gray value at the target location has a larger ratio to the sum of the pixel gray levels in the surrounding small area.Local probability distribution is used to detect and extract the characteristics of the target.Since the target g roup loses the typical characteristics of a single target when the occlusion overlap occurs,the single target is not detected at this time,and the overlapped part of the target is detected together,that is,the single The pixel cluster formed by the target is regarded as a set of the same target,and the obtained detection result provides a measurement set for the tracking and filtering of the target.(3)Target tracking method based on Gaussian mixture PHD filter.Since the number of targets in the grou p target is not fixed during the release process,the target overlap and occlusion will occur.Aiming at the data association problem caused by the traditional target tracking method,the multi-target tracking theory based on random finite set is applied,and the recursive probability is adopted.Assuming density,real-time estimation of the number and state of group targets is achieved.However,due to the existence of set-valued integrals,it is difficult to calculate,and because the analytical form of t he multi-objective Bayesian formula is difficult to solve,the Gaussian mixture PHD method is used to achieve the group target set Same tracking.