D-S证据理论改进算法提高水下目标识别准确性  

Algorithm Improvement of D-S Evidence Theory in Submarine Target Recognition

作　　者：刘标 许腾[1] 李光 

机构地区：[1]海军指挥学院,江苏南京210000 [2]海军士官学校,安徽蚌埠233012

出　　处：《现代防御技术》2018年第1期120-124,155,共6页Modern Defence Technology

摘　　要：D-S证据理论在目标识别中有着广泛的应用,是较好的数据融合方法之一,但在证据高度冲突时,会产生有悖常理的结果。针对此问题,引入传感器的可信度,提出了一种新的改进算法。该算法弥补了D-S证据理论所存在的不足,比其他改进算法融合效果好,此算法应用于水下目标识别,计算结果表明提高了水下目标识别的准确性和有效性。D-S evidence theory is widely used in target recognition. It is an effective method of multi-sensor data fusion. However,it involves counter-intuitive behaviors when the high conflict information exists. In view of the problem,the reliability of sensor is introduced and the new algorithm is proposed. The algorithm makes up for the shortcomings of D-S evidence theory,and has better fusion effect than other improved algorithms. The algorithm is applied in submarine target recognition and the calculation result indicates that the algorithm is effective and correct.

关 键 词：数据融合 多传感器 目标识别 证据理论 改进算法 可信度 

分 类 号：E967.3[军事—军事通信学] TP212[自动化与计算机技术—检测技术与自动化装置]