一种复杂环境下的多传感器目标识别系统  

A Multi-sensor Target Recognition System in the Complex Interference Environment

作　　者：蒲书缙 杨莘元[1] 赵忠凯[1] 赵坤[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《兵工学报》2006年第3期418-421,共4页Acta Armamentarii

摘　　要：对于复杂环境下的多传感器目标识别,其环境因素多变、处理数据量庞大,若采用传统的识别方案,系统的可靠性和抗干扰性不理想。为了解决这些问题,提出了一种用于复杂环境下的神经网络分类器组合结构的多传感器目标识别系统,并通过几种融合方法的仿真实验表明该系统是可行的,能有效地提高系统识别率及鲁棒性。For multi-sensor target recognition in a complex interference environment, the environment is variational and the data processed is massive. If conventional methods are used, reliability and robustness of the system hardly meet the requirements. To solve these problems, a multi-sensor target recognition system was presented based on the ensemble of neural network classifiers in the complex interference environment. Several fusion methods were simulated, and the results show that the system is practicable and it can effectively enhance the efficiency and robustness of target recognition system.

关 键 词：信息处理技术 多传感器数据融合 目标识别 神经网络 可信度 决策融合 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]