User Tools

Site Tools


Mid-Infrared Explosives Detection and Imaging

Mid-infrared radiation is used to detect and image trace amounts of explosives on varying substrates. e.g. car doors, travel bags, and cloth.


Detection of trace amount of explosives is a primary concern of defense administrations around the world. Ideally this detection would be done in a way to enable quick and discreet identification of these explosive materials while suppressing false positives.


Organic compounds, including many common explosives, exhibit spectral “fingerprints” in the mid-infrared(MIR) region of the electromagnetic spectra. This allows identification of the compound by use of MIR spectroscopic techniques. Utilizing one of the regions of the electromagnetic spectra where atmospheric absorption is low, a so called atmospheric window, a practical stand-off detection system can be realized. This system works by conducting a diffuse reflectance measurement of the object of interest. By comparing the measured spectra with a reference spectra identification of various explosives can be accomplished.

The first step in building a detection system is identification of the spectral fingerprints. Figure 1 shows the reflectance spectra of Semtex measured from a plastic background. However, these spectra vary largely over different substrate/explosive combinations making the construction of a spectral library containing every combination difficult/impractical. Instead the approach taken here is to use Principal Component Analysis(PCA) to identify the important spectral features and compare these to known spectra instead. Figure 2 shows the 2nd derivative first principal components of Semtex and C4 measured on the same plastic background. The 2nd derivative is used in order to remove any affects due to intensity variations. What can be seen is that the majority of the variation in both cases can be attributed to the peak at 1260 cm^{-1}. This peak has been attributed to Research Department explosive(RDX) as it is the explosive component in both C4 and Semtex, this agrees with spectra found in the literature. Finally, figure 3 shows an example of using the first principal components in order to identify explosives. This plot was generated by taking spectra of a spot on a substrate and computing the 2nd derivatives 1st principal component. By plotting the score of the first principal component one can identify the areas that have explosives. In the case shown here all positive values correspond to no explosives and all negative values correspond to areas with explosives. No false positives nor negatives were observed.

Figure 1 Figure 2 Figure 3

Future Work

Using the spectral fingerprints identified by PCA work is now being conducted to build the detection system.

projects/mid-infrared_explosives_detection_and_imaging.txt · Last modified: 2016/01/14 13:11 by dbenirsc