多传感器数据融合模型综述

多传感器数据融合的关键问题是模型设计，数据融合模型主要包括功能、结构和数学模型。根据数据融合功能层次性和信息流通方式及传输形式，把数据融合功能模型分为五级，即检测级、跟踪级、属性级、态势级和威胁级。这种划分方法更有利于数据融合技术的研究。在此基础上，详细讨论了前三个级别的系统结构模型、数据融合系统的组成、相关数据流、系统与外部环境的界面及各种结构模型的优缺点和应用范围，并同时介绍了有关的融合算法。

Survey of multisensor data fusion models

The key problem of multisensor data fusion is designing of model. Data fusionmodels mainly contain function models, architectural models and mathematical models. Thispaper classifies the function models of data fusion into five levels according to the functionalhierarchy of fusion and mode of data process flows, i. e., detection level, tracking level, attribute level, situation level and threat level. This classification method is particularlyfavourable for the study of data fusion algorithms. On the basis of the above classification,the system architectural models in the preceding three levels and their advantages and disadvantages for both military and nonmilitary applications are disccussed and compared in detail. an introduction of the benifits of all fusion fashion and data fusion applications is alsoprovided, in the meantime relative fusion algorithms are introduced. Finally, a conclusion ofthe paper is given.