多传感器信息融合分类识别系统中冗余知识判定的关系代数方法

A Relational Algebra Method to be Used to Decide Redundant Knowledge in Classification and Recognition Systems based on Multisensor Data Fusion

在基于多源信息融合的对象分类识别系统中，信源／ 传感器数量的增加往往使数据处理量急剧增长，不仅占用大量计算资源，提高运行成本，而且降低了系统可贵的实时性。本文提出用关系代数方法判定由信源传感器提供的知识有无冗余，从而精简和优化信源／ 传感器资源的配置和组合。

In a classification and recognition systems of object based on multisensor data fusion, the increase of information resources and sensors will cause a speedy growth of data processing. It will not only occupy a great deal of computing resource and raise operating costs, but also lower valued reality of system. The thesis proposes a relational algebra method that can be used to decide whether the knowledge provided by resources or sensors is still redundant, and to refine or optimize both configurations and combizations of resources and sensors.