摘要
针对多个传感器对多个特性指标进行测量实验的数据融合问题,从多元统计理论角度提出了一种新的多传感器数据的融合算法.该方法采用欧氏距离定义了距离矩阵,利用最小距离聚类法确定各传感器融合的次序,可以克服以往方法中关系矩阵的主观影响,提高数据融合结果的客观性.通过试验数据的分析,表明该算法简单,可以避免极端、有效数据的损失,具有较高的精度.
Due to data fusion of multi-sensors experiment on many characteristic indexes, a new fusion arithmetic for multi-sensors data is brought forward from the multivariate statistical theory. The arithmetic defines the distance matrix by the Euclidean distance and determines the orders of sensor fusion according to the minimum distance clustering. It may overcome the influence of subjective factor for the relationship matrix and improve the objectivity of data fusion. The analysis for the experiment data shows that the method is very simple, can avoid losing the important extreme data and effective data, and has better accuracy.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2008年第5期131-135,共5页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(10626029)
江西省自然科学基金(0611082)
江西省教育厅科技项目(GJJ08350)
关键词
多传感器
数据融合
特征指标
最小距离聚类法
multi-sensors
data fusion
characteristic index
minimum distance clustering