摘要
随着防空作战环境的日益复杂,快速准确地进行空中目标识别显得尤为重要。多元信息融合为目标识别提供了一个新的思路,信息融合可在像素级、特征级及决策级3个层次上进行。基于证据推理的决策级融合以其在自动目标识别中特有的优势而成为研究的热点。为了解决Dempster组合规则在高冲突证据下存在的问题,对组合规则的思想进行了分析;讨论了针对该组合规则的两大主流修正方向:基于修正证据源的改进措施和基于修正Dempster组合规则的方法一些改进措施;针对空中目标识别问题,提出了综合运用各种改进思想的融合方法,给出了识别过程的流程图,算例结果表明本文方法的有效性。
With the environment in aerial defending becoming increasingly complicated, it is very important to recognize the target fleetly and exactly, Multi - data fusion provides a new idea in target recognition. Information fusion can be performed at pixel level, feature level or decision level. Dempster＇s rule plays a potential role in auto - recognition of target, to solve the problems in the application of the evidence Theory, this paper analyzes the princi- ple of Dempster＇s rule and then discusses its limitations and arnendatory measures, including amending evidence source and modifying fusion rule. Finally, the paper proposes a general method for air target identification and presents the flow chart of the identification process. Through simulation and the comparison of it with other algorithms, the results show that the proposed algorithm is more efficient and feasible.
出处
《空军工程大学学报（自然科学版）》
CSCD
北大核心
2008年第4期50-53,共4页
Journal of Air Force Engineering University(Natural Science Edition)
