一种用于电视跟踪系统中的多模式数据融合方法  

A multimode data fusion method used for TV tracking system

作　　者：王军宁[1] 龚莹[1] 李波[1] 

机构地区：[1]西安电子科技大学ISN国家重点实验室,陕西西安710071

出　　处：《光电工程》2004年第5期28-31,共4页Opto-Electronic Engineering

基　　金：国家自然科学基金资助(No.60002007)

摘　　要：基于单传感器多模式数据融合的思想,设计了一种联合卡尔曼滤波器。通过卡尔曼滤波分 散处理多个模式跟踪结果数据,再将处理结果进行全局融合得出跟踪最终结果。仿真和室内实验都表明,这种方法用于实时电视跟踪系统中可有效提高跟踪精度,融合后的均方误差是融合前的0.05倍,具有一定的容错能力,运算量小,且易于实时实现。A federated Kalman filter is designed based on multimode data fusion for single sensor. The final tracking results can be obtained through dispersion processing multimode tracking data with Kalman filter and performing global fusion for the processed results. Simulation and indoor experiments show that this method can effectively improve the tracking accuracy when it is applied to TV tracking system. RMS error of this method after data fusion is 0.05 times as that before data fusion. It has certain fault-tolerant capability. Its computation is small and its operation can be easily performed at real-time.

关 键 词：数据融合 多模式融合 电视跟踪系统 卡尔曼滤波 

分 类 号：V556.5[航空宇航科学与技术—人机与环境工程]