多分辨率多传感器动态数据的融合和应用

基于线性均方估计建立了q 叉树的多尺度自回归模型. 该模型应用于分布式多分辨率多
传感器进行建模, 可实现多分辨率多传感器的动态数据融合. 将该算法用于高精度划线切割机器人系统中多传感器的建模, 实现了型钢划线切割过程中型钢边缘的动态检测. 实验结果表明, 多分辨率多传感器数据融合可消除噪声的干扰, 提高检测系统的测量精度和整个加工系统的加工精度.
关键词: 多传感器; 动态数据融合; q 叉树; 边缘检测
Abs tra c t: A class of m u lt iscale au to regression model based on qth2o rder t ree w as in t roduced and the algo2 rithm is based on the linear m ean square est im aiton. The model w as app lied to deal w ith the p rob lem of dist ribu ted m u lt ireso lu t ion dynam ic data fu sion. The model w as also app lied to model the m u lt isen so r in h igh p recision m ark ing2cu t t ing robo t system and to do the dynam ic m easu rem en t of p rof iled bars’ edge in
the m ark ing and cu t t ing p rocess. The experim en tal resu lt show s that the m u lt ireso lu t ion m u lt isen so r data fu sion can deno ise and enhance the p recision of m easu ring and m ach in ing of the system.
Ke y w o rds: m u lt isen so r; dynam ic data fu sion; qth2o rder t ree; edge m easu rem en t