基于小波变换的白酒检测电子鼻降噪方法研究

Research on Denoising Method of Electronic Nose for Liquor Detection Based on Wavelet Transform

电子鼻在检测白酒气体时,数据会受到环境以及电子鼻自身噪声的干扰。为降低各种噪声对数据产生的影响,文中在小波变换降噪的基础上,提出一种改进的小波阈值函数算法。利用仿真验证了改进后的算法的有效性,并通过信噪比和均方根误差对改进后算法的降噪效果进行定量评价。然后,将改进的阈值函数小波降噪算法应用到白酒检测电子鼻中,并与软、硬阈值函数小波降噪算法和中值滤波加Savitzky-Golay滤波降噪算法进行对比。仿真和实验的结果表明,改进的阈值函数小波降噪方法可以在尽可能保持原有信号完整性的基础上,对噪声进行有效地过滤,利于白酒种类的分类。

When the electronic nose(e-nose)is used to detect liquor gas,the data is often disturbed by the noise generated by the e-nose itself and the environment.In order to reduce the influence of various noises on data,this paper proposed an improved wavelet threshold function algorithm based on wavelet transform denoising.The effectiveness of the improved algorithm was verified by simulation,and the noise reduction effect of the improved algorithm was quantitatively evaluated by signal-to-noise ratio and root mean square error.Then,the improved threshold function wavelet denoising algorithm was applied to the liquor detection e-nose,and compared with the soft and hard threshold function wavelet denoising algorithm and median filter plus Savitzky-Golay filtering denoising algorithm.The results of simulation and experiment show that the improved threshold function wavelet denoising method can effectively filter the noise on the basis of maintaining the original signal integrity as far as possible,which is conducive to the classification of liquor types.