基于正常行为的案例推理在伪装船只识别中的应用

Application of CBR Based on Normal Behavior in Camouflage Vessel Identification

船只识别是一个多影响因素的问题,目前的船只识别方法,普遍采取基于成像原理的方法,利用船只外形的差异,进行识别。对于伪装船只,其外形与普通船只差别较小,利用成像技术对其进行识别比较困难。提出了一种基于船只行为的识别方法,将邻域粗糙集、案例推理与聚类分析引入船只识别的过程中,利用邻域粗糙集对案例进行属性简约,提取船只行为案例关键属性,再对案例库进行案例聚类,以提高案例的检索效率。利用提出的推理模型,充分发掘船只动作数据所包含的信息,能以较高准确率实现对伪装船只的快速识别。

Ship identification is a problem of many influencing factors.At present,ship identification generally adopt the method based on the imaging,and the differences in the shape of ships are used to implement the identification.Because of the tiny differences between camouflaged ships and ordinary ships,it is difficult to identify them using imaging technology.Introducing neighborhood rough set,case reasoning and case clustering into the process of ship identification,a method based on ship behavior is proposed.Neighborhood rough set is used to simplify the attributes of cases and extract the key attributes of ship behavior,after that,case cluster is used to improve the efficiency of case retrieval.Using the proposed reasoning model to fully explore the informa?tion contained in the ship's movement data,camouflaged ships can be rapidly recognised with high accuracy.