数据融合技术及其应用

数据融合是对多传感器目标信息按时序和准则自动分析和综合处理得出最终结论或决策的技术，包括多传感器和多信息输入、合成规则、表示形式等，分象素／特征／决策三级。随机融合包括经典推理法、卡尔曼滤波法和专家系统法等，用于三级融合各阶段。人工智能融合有模糊集合、神经网络、小波分析等，适用高层次融合，数据融合系统由传感器、处理机及相关硬软件组成，按结构分为并联／串联／混合／网络型，按控制开关分为开／闭环型，按战术应用分为集中／分布式，按融合规模分为单／多平台。

Data Fusion Technique and Its Application

Date fusion is a technique that analyzes automatically and process synthetically the multi-sensors goal information based on time and rules to get the conclusion or decision. This technique includes multi-sensors input, multi-information input, composing rules and expressing form. There are three levels and two methods: the levels of pixel, character and decision. Randomizing method includes classical organon, Kalman filter and expert system, which can be used in each level. AI method includes fuzzy aggregate, nerval net and little wave analysis, which can be used in higher level. Date fusion system consists of sensors, disposing machines and operation software and hardware. Date fusion is split into kinds of parallel connection/series/commix/network type structure, open/closed loop type structure according to control switch, and centralizing/distributing structure according to tactics, single/multi-platform structure according to data fusion size.