Non-invasive measurement study of human blood alcohol concentration based on NIR dynamic spectrum

Abstract

By collecting the NIR spectroscopy of human finger end artery blood, the dynamic spectrum data can be obtained. Partial least square (PLS) regression was used to establish the calibration model of dynamic spectrum data and volunteers' biochemical analysis alcohol concentration values, and then the human blood alcohol concentration can be predicted. 104 finger end dynamic spectral data, obtained from 104 volunteers after drinking, the calibration set contains 80 swatches, and the prediction set contains 24 swatches, for calibration set and prediction set, the correlation coefficient are 0.9672 and 0.9384. For prediction set, the maximum relative error is 9%, the minimum relative error is 0.6%, and the average relative error is 3.26%. The measurement results show that NIR dynamic spectrum method can be applied to realize the non-invasive measurement of alcohol concentration in human blood. It can be a very desirable non-invasive measurement method for alcohol concentration in human blood.