基于小波变换的体内外酒精含量近红外光谱检测与分析  

Detection and Analysis of Alcohol Near-Infrared Spectrum in Vitro and Vivo Based on Wavelet Transform

作　　者：罗斯特[1] 李增勇[1] 张明[2] 陈国强[1] 

机构地区：[1]山东大学机械工程学院,山东济南250061 [2]香港理工大学医疗资讯科技学系

出　　处：《光谱学与光谱分析》2012年第6期1541-1546,共6页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(30870652;81071223);山东省自然科学基金项目(ZR2010HM024);山东大学自主创新基金自然科学类专项(2010JQ007)资助

摘　　要：应用小波分析对体外和体内的酒精近红外光谱信号进行去噪分析,通过体外光谱分析确定酒精吸收峰特征范围,为体内近红外光谱分析确定有效区间。软阈值和硬阈值下,分别采用缺省阈值、Birge-Massart阈值和最大最小值阈值,比较酒精光谱去噪,信噪比(signal noise ratio,SNR)和均方根误差(root mean square error,RMSE)去噪效果。结果表明:缺省硬阈值方法对酒精近红外光谱去噪的效果较好;小波变换可以有效去除酒精近红外光谱的噪声,提高信噪比,保留有用真实信号。在不同的酒精浓度下,去噪后的近红外光谱能够较好的显示浓度变化规律。小波分析在近红外光谱法对人体酒精无创检测及定量分析方面有较好的应用前景。The near-infrared spectroscopy（NIRS） signals of alcohol in vivo are always contaminated by noise.In the present study,wavelet analysis was used to eliminate noise and thereby detecting the NIRS signals of alcohol in vivo.In soft and hard threshold function,the spectral signals were de-noised by default threshold,Birge-Massart threshold and minimax threshold respectively.Signal noise ratio（SNR） and root mean square error（RMSE） method were used to evaluate the effects of the de-noising.The results show that the default threshold de-noising has the best effects.Therefore,the default threshold de-noising was chosen to perform de-noising analysis in vivo and in vitro.Our result shows that the wavelet transform de-noising is effective in removing noise from NRS signals of alcohol in vivo.With different alcohol concentration,the de-noising spectrua can show evident absorption peaks.Wavelet analysis is an effective method in the non-invasive alcohol testing and quantitative analysis.

关 键 词：小波变换 近红外光谱 酒精 去噪 阈值 

分 类 号：R318.5[医药卫生—生物医学工程]