人体运动实时捕捉设备传输数据的滤波与处理  

Filter and processing of data transferred from real-time human motion capture device

作　　者：高申玉[1] 刘金刚[1] 

机构地区：[1]中科院计算所&首都师范大学计算机科学联合研究院

出　　处：《计算机工程与设计》2006年第15期2715-2717,2773,共4页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2001AA114201)。

摘　　要：简介人体运动实时捕捉设备的工作原理,对设备探测并传输来的数据进行特性分析,根据分析结论,提出了干扰数据识别与修复、纵向平滑滤波、首帧收敛化处理和值域限定等滤波和处理方法,并详细阐述了这些方法的综合运用。系统使用这些方法,经运行调试后,传输数据稳定,系统对Marker点的探测精度也符合系统设计的要求(6mm),为实现各点坐标计算提供了必要条件。The principle ofreal-time human motion capture device is introduced briefly, the characteristics ofthe detected data that transferred from the device are analysed. Based on the analysis, some filters and disposal methods are put forward, such as identifying and repairing disturbing data, lengthways smoothed filter, head-flame data convergence disposal, and number field restriction. The synthetic use ofthese methoda is expatiated. Byusingthesemethoda, the system run well and the transferred data are stable, the devices＇ detecting precision measures up the designing index （≤6 mm）, and gives essential conditions for coordinates calculating.

关 键 词：运动捕捉 数据处理 纵向平滑滤波 首帧收敛化处理 值域限定 

分 类 号：TP911.72[自动化与计算机技术]