基于D-S证据理论的多特征数据融合算法  

The Algorithm of Multi-feature Data Fusion based on D-S Theory of Evidence

作　　者：宋建勋[1,2] 张进[1,2] 吴钦章[1] 

机构地区：[1]中国科学院光电技术研究所,成都610209 [2]中国科学院研究生院,北京100039

出　　处：《火力与指挥控制》2010年第7期96-98,108,共4页Fire Control & Command Control

摘　　要：Dempster-Shafer证据理论是不确定推理的一种重要方法,提供了一定程度的不确定性,可以指定给相互重叠或互不相容的命题,然后通过Dempster组合规则将不确定性信息在新证据中进行重新分布,该理论在许多方面都得到了广泛的应用。将来自图像传感器的多种图像特征,通过D-S证据理论将这些特征信息进行融合,并应用于目标的识别。实验结果表明D-S证据理论用于多特征数据融合的目标识别算法是有效的,基于该理论的多特征数据融合具有广阔的应用前景。The Dempster-Shafer theory of evidence is an important tool of uncertainty reasoning, and it provides the method of uncertainty, and the evidence may be designated to the overlapping and incompatible proposition, and then the information of uncertainty is distributed in the new evidence according to Dempster combination rule, and the theory is widely used in many fields. In this paper, multiple features are got from the sensor of image, and they are fused to a integrated feature, then the feature is used in target recognition. It is shown that the method is effective, and it will have a bright future.

关 键 词：D-S证据理论 多特征 数据融合 目标识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]