战场侦察数据在频域上的异常分析

Anomaly Analysis of Battlefield Reconnaissance Data in Frequency Domain

针对战场在对固定区域进行观察和监控后,获得大量的数据在频域上的异常数据分析处理问题,提出了一种将二次型距离与K-means聚类算法相结合的改进的K-means聚类算法以适用于所研究的场景,从而更准确地预警敌方战略的变化,以辅助指挥官决策。仿真结果显示该方法可以良好地分析出战场固定区域内各频段上目标数量变化的异常情况。

Aiming at the problem of analyzing and processing massive data obtained from observation and monitoring of fixedarea in the battlefield,an improved K-means clustering algorithm which combines quadratic distance and K-means clustering algo-rithm is proposed to suit the scenario we studied so as to more accurately warn the enemy＇s strategic changes to assist commander de-cision-making. The simulation results show that this method can well analyze the anomaly of the change of target quantity in each fre-quency band in the fixed area of the battlefield.