基于DSP的人体手势动作实时识别系统

针对人体表面肌电信号（SEMG）的非平稳性、小波包变换系数维数过高和识别率低的问题，设计了基于DSP处理器TMS320VC5502硬件平台的便携式人体手势动作实时识别系统，并提出了一种小波包主元分析（WPPCA）和线性判别分析（LDA）相结合的表面肌电信号动作特征识别新方法。实验结果表明，该方法能够将小波包系数矩阵由16维降到4维，并且对前臂的握拳、展拳、手腕内翻和手腕外翻4种动作模式的平均正确识别率达99．5％，与传统的小波包变换相比有较高的识别率。...

Full description

Saved in:
Bibliographic Details
Published in电子技术应用 Vol. 40; no. 7; pp. 75 - 78
Main Author 秦勤 李艳玮
Format Journal Article
LanguageChinese
Published 河南工程学院计算机学院,河南郑州,451191 2014
Subjects
主元分析
动作识别
线性判别分析
表面肌电信号
设备控制
动作识别
device control
表面肌电信号
action recognition
设备控制
主元分析
SEMG
linear discriminant analysis
线性判别分析
principal component analysis
Online AccessGet full text
ISSN0258-7998

Cover

More Information
Summary:针对人体表面肌电信号（SEMG）的非平稳性、小波包变换系数维数过高和识别率低的问题，设计了基于DSP处理器TMS320VC5502硬件平台的便携式人体手势动作实时识别系统，并提出了一种小波包主元分析（WPPCA）和线性判别分析（LDA）相结合的表面肌电信号动作特征识别新方法。实验结果表明，该方法能够将小波包系数矩阵由16维降到4维，并且对前臂的握拳、展拳、手腕内翻和手腕外翻4种动作模式的平均正确识别率达99．5％，与传统的小波包变换相比有较高的识别率。
Bibliography:In view of surface electromyography （SEMG） non-stationary characteristics, higher dimension of wavelet packet transform coefficients and low recognition rate, a portable human gestures real-time recognition system based on TMS320VC5502 DSP processor hardware platform is designed ,and a new method of SEMG signal feature recognition which is combined wavelet packet principal component analysis （WPPCA） and linear discriminant analysis （LDA） is proposed. Experiments show that this method can reduce the wavelet packet coefficient matrix of surface EMG signal from 16 dimensional to 4 dimensional, and successfully identify four kinds of motions such as hand grasping, hand opening, radial flexion and ulnar flexion, and action recognition rate is up to 99.5%, which has a higher recognition rate comparing to the traditional wavelet packet transform, this algorithm.
11-2305/TN
principal component analysis; action recognition; SEMG; linear diseriminant analysis; device control
Qin Qin, Li Yanwei （College of Computer, Henan
ISSN:0258-7998