多分辨率小波变换在表面肌电信号模式识别中的应用

The Multi-resolution Wavelet Transform Applying in Pattern Recognition of EMG

采用一种多分辨率小波变换来进行表面肌电信号的模式识别。该方法选择各个尺度下的小波系数幅值的一对最大最小值作为特征向量,对表面肌电信号作5尺度小波分解。对内旋动作、外旋动作、握拳动作、展拳动作的多尺度分解结果证明,该方法提取的特征信号用于识别时比传统方法的分类精度高。

We can recognize EMG with a multi-resolution wavelet transform based on pattern recognition in this paper.The method is decomposing the EMG for 5-scale wavelet by selecting a minimum and maximum value of amplitude of wavelet coefficients in various classes of the feature vectors.The Multi-scale decomposition of the internal rotation motion,fist action,exhibition boxing showed that the method had higher precision than traditional one when it was used to Collect Signature for signal recognition.