基于选择因子的皮肤电反应信号小波去噪方法  

GSR Signal Denoising Method Via Selection Factor and Wavelet

作　　者：刘名扬[1] 张晓寒[1] 范頔[1] 宫晓鹏[1] 

机构地区：[1]吉林大学仪器科学与电气工程学院,长春130012

出　　处：《科学技术与工程》2017年第15期238-241,共4页Science Technology and Engineering

基　　金：吉林大学校级"大学生创新创业训练计划"创新训练项目(2015650950);吉林省科技厅国际合作项目(20160414019GH)资助

摘　　要：在皮肤电反应(galvanic skin response,GSR)信号采集过程中,由于环境等因素不可避免地会引入噪声,需要在进行模式识别前对采集信号进行去噪处理。利用MATLAB研究不同小波组合对GSR信号的去噪效果;并使用信噪比、均方差为评价指标,提出了一种基于选择因子的GSR小波去噪方法。利用采集得到的GSR信号,选择因子选择出来的小波组合,对GSR信号去噪处理。实验中采用db5小波对信号进行5层分解,在sqtwolog或heursure阈值选择方法下对信号进行的去噪处理,从而达到了理想的滤波效果。In the process of Galvanic Skin Response（ GSR） signal acquisition, various of reasons such as envi-ronmental factors inevitably will introduced noise. Before pattern recognition, the noise be should removed which superposed upon the collected signals. MATLAB has been used to study GSR signal denoising effect with different wavelet combination, and signal-to-noise ratio and the mean square error as the evaluation index were used. A de-noise method was putting forward by wavelet based on selective factor to get the good combination of wavelet GSR signal denoising. In the experiment, the db5 wavelet function of decomposition for 5 layer and the threshold selec-tion method of sqtwolog or heursure to denoise the signal were used,and the ideal filtering effect was got.

关 键 词：小波去噪 GSR信号 选择因子 滤波评价 

分 类 号：TN911.7[电子电信—通信与信息系统]