Techniques for quantifying chemicals concealed behind clothing using near infrared spectroscopy

Abstract

The detection of specific chemicals when concealed behind a layer of clothing is reported using near infrared (NIR) spectroscopy. It is found that concealment modifies the spectrum of a particular chemical when recorded at stand-off ranges of three meters in a diffuse reflection experiment. Chemometric analysis of the spectra has been performed with neural network-based pattern recognition/classification to deal with this problem. Neural networks help to overcome nonlinearities within the calibration/training dataset, affording more robust modelling. The work has been shown to both allow detection of specific chemicals concealed behind a single intervening layer of fabric material, and to estimate the concentration of hydrogen peroxide using partial least squares regression (PLSR).