基于MEMS传感器的三维空间运动轨迹提取方法  

THREE-DIMENSIONAL SPACE TRAJECTORY EXTRACTION METHOD BASED ON MEMS SENSOR

作　　者：仇立杰[1] 彭四伟[1] 

机构地区：[1]北京化工大学信息科学与技术学院,北京100029

出　　处：《计算机应用与软件》2016年第9期292-295,共4页Computer Applications and Software

摘　　要：基于MEMS传感器的手势识别的一般方法是通过对每个手势进行运动学分析找出每个手势的加速度特征,以此作为识别手势的特征值。这表明该方法需要人工设计特征值,不利于用户的自定义操作。针对这种情况,提出一种空间定位方法。首先获取传感器数据并发送到PC端,并在PC端进行积分处理,转化成空间数据,从而记录手势的运动轨迹。最后结果表明,该方法可以成功记录手部的运动轨迹,不用对每个手势单独进行运动学分析,使得用户在后续的研究中可以实现自定义手势。The general approaches of MEMS sensor-based gesture recognition are to find the acceleration characteristics of each gesture through kinematic analysis on every gesture, and then use them as the characteristic values of gesture recognition. This indicates that the methods require manually designed characteristic values, which is not conducive to user＇ s self-defined operations. In view of this, we propose a space positioning method. First, it acquires the sensor data and sends it to PC side, and then the sensor data is made the integration processing on PC side to transform it into spatial data, thereby the motion trajectory of the gesture is recorded. Final results show that the method can successfully record the motion trajectory of hand, and avoid the single kinematic analysis on every gesture, this enables the user to be able to realise self-deflning the gesture in subsequent studies.

关 键 词：传感器数据 空间数据 手势识别 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]