基于贝叶斯网络的多传感器目标识别算法研究

Research on Multisensor Target Recognition Algorithm Based on Bayesian Networks

基于贝叶斯网络能够组合多种证据进行不确定性表达和推理的特点,提出以贝叶斯网络为基本结构的目标融合识别模型.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.首先对各传感器的数据分别进行融合,然后应用贝叶斯网络推理算法对多种传感器融合结果进行融合计算,最后根据假定变量各状态的概率取值来判断目标平台类型.仿真结果证明了该方法直观、形象,计算速度快,降低了实用的复杂度,提高了目标识别的可靠性.

Due to the character of Bayesian network which can express and infer on uncertainty evidences. The target recognition modeling based on bayesian network has been brought out. The bayesian network struction of target fusion racognition has been built based on the inference rules of air target recognition. The data fusion of diferent sensors was delt first, their fusion results were fused again with bayesian network inference algorithm and the final fusion result can be gained according to the probability values of different variable states. The simulation example showed that this method is simple and direct in computation in real time,reduces the complexity in application and approves the reliaability in recognition.