基于模糊贝叶斯网络的空中目标多传感器融合识别研究  

Aerial Target Fused Recognition Based on Fuzzy Bayesian Network Model

作　　者：刘海燕[1,2] 陈红林[2] 史志富[1] 梁华强[2] 

机构地区：[1]第二炮兵工程学院,西安710025 [2]西北工业大学,西安710072

出　　处：《电光与控制》2009年第3期37-41,共5页Electronics Optics & Control

摘　　要：为了对充满不确定性与模糊性的空中目标识别数据进行处理,提高空中目标多传感器融合的准确性和可靠性,提出以模糊贝叶斯网络为基本结构的多传感器数据融合模型。该模型能够对清晰连续变量通过模糊化和去模糊化操作变换成离散变量,而且基于模糊贝叶斯网络的建模方法能够组合多种证据进行不确定性表达和推理。通过详细分析空中目标识别的推理规则,建立了空中目标识别的贝叶斯网络拓扑结构,提出了贝叶斯推理算法对多种证据进行融合计算的模型。识别实例表明该模型能够融合不同信息源的数据,有效地提高空中目标识别的效率。For dealing with aerial target recognition data with uncertainty and fuzzyness, and improving the precision and reliability of multi-sensor fusion for aerial target, a multi-sensor data fusion model was put forward based on fuzzy Bayesian network structure. Fuzzy Bayesian networks can transform a continuous variable into a discrete variable through fuzzification and defuzzification. Fuzzy Bayesian network was an efficient way for combining evidences from different sources and reasoning under uncertainty. The inference rule for aerial target recognition was analyzed, and a Bayesian network topological structure was built up for aerial target recognition. The different data sources have been fused with Bayesian probability reasoning. The simulation examples showed that the method can fuse the different resources and can improve the efficiency of aerial target recognition.

关 键 词：空中目标识别 模糊贝叶斯网络 多传感器融合 贝叶斯概率推理 

分 类 号：V233.7[航空宇航科学与技术—航空宇航推进理论与工程] TP11[自动化与计算机技术—控制理论与控制工程]