基于目标分类识别的跟踪方法研究

Research on tracking method based on target classification and recognition

针对空中未知目标的精确跟踪问题,采用基于目标分类识别的方法,针对不同类型的目标选用不同运动模型,以提高对未知目标的跟踪精度。针对3种空中目标:战斗机、民航机和直升机,首先为3种目标设计准确的运动模型;用贝叶斯推理基于目标状态及信号对目标进行分类识别,根据分类识别结果,选择匹配的运动模型进行跟踪。仿真结果表明,相较于标准交互多模型IMM算法,所提方法能够使目标跟踪的误差减小10%左右。说明,利用目标的分类识别的结果,对目标运动状态的精确建模,能够有效提高未知目标的跟踪精度。

For the precise tracking problem of unknown targets in the air,the method based on target classification and recognition is used to select different motion models for different types of targets to improve the tracking accuracy of unknown targets.The article considers three air targets:fighters,civil aircraft and helicopters.Firstly,an accurate motion model is designed for the three targets.Bayesian inference is used to classify and identify the target based on the target state and signal.According to the classification recognition result,the matching motion model is selected for tracking.The simulation results show that compared with the standard Interacting Multiple Model(IMM)algorithm,this method can reduce the error of target tracking by about 10%.It shows that the accurate modeling of the target motion state can effectively improve the tracking accuracy of the unknown target by using the result of the classification recognition of the target.