Spectral Analysis for the Detection of Explosives with Differential Reflectometry Seniha Esen Yuksel Dept. of Computer, Information Science and Engineering University of Florida Gainesville, FL seyuksel@cise.ufl.edu Thierry Dubroca Dept. of Material Science and Engineering University of Florida Gainesville, FL dubroca@ufl.edu Rolf E. Hummel Dept. of Material Science and Engineering University of Florida Gainesville, FL rhumm@mse.ufl.edu Paul D. Gader Dept. of Computer, Information Science and Engineering University of Florida Gainesville, FL pgader@cise.ufl.edu ABSTRACT For explosive detection purposes, it is assumed that the per- son preparing or carrying the explosive will inadvertently contaminate him/herself or the exterior of the package. To detect such traces of explosive materials, we show the use of differential reflectometry (DR) as an alternative system to the existing techniques. With DR, explosives show char- acteristic behaviours at specific wavelengths, for example, spectra of TNT shows a sudden decrease at 420 nm. To detect these behaviours, principle component analysis was performed to reduce the dimensionality of the data, and a support vector machine classifier was trained to identify TNT. With a 10-fold classification on 10000 non-TNT and 1935 TNT pixels, we achieved 0.3% false alarm rate at 75% true positive rate. In this study, we outline the operation of the DR system, show the unique signatures of explosives when viewed with DR, and report the detection rates with support vector machine classifiers. Keywords Explosive detection, hyper-spectral imaging, spectroscopy, differential reflectometry, classification, SVM, TNT. 1. INTRODUCTION One of the current explosive detection methods that is being used in the U.S. airports is ion mobility spectroscopy (IMS). In this system, a technician wipes a cloth over a piece of lug- gage which in turn is placed into the IMS device that heats this swab to vaporize the particles to be investigated. The particles are ionized by electron bombardment and then ac- celerated in an electric field. Different ions have different speeds. Using the speed information the presence of nitro- gen can be deduced [9]. However, IMS has a number of disadvantages; for example, the time required for wiping an entire bag does not allow to investigate all pieces of luggage inside and outside which means that the technician has to make some judgment which bag to investigate. On the other hand, if all bags are screened, the slow process leads to long lines at the airports. Also, IMS cannot differentiate between explosives and other nitrogen containing substances such as fertilizer, cosmetics and certain polymers which often lead to false positives [5]. Most of all however, IMS needs the involvement of an operator which is costly. For detection of explosives on humans, the so-called“puffer” has been tried which involves blowing air in a closed cell to dislodge possible explosives, then vacuum the air in this cell to direct it into an IMS device, as just described [5, 6]. However, it has been shown that explosive molecules can stick to surfaces with relatively high binding energies [4, 6] which may preclude the release of the explosive vapor. In order to alleviate these problems differential reflectome- try (DR) has been developed [3] at the University of Florida. This technique, also called as differential reflection spec- troscopy, measures the differential reflection from materials at multiple wavelengths. With DR, explosives show signa- ture patterns at specific wavelengths, and with the design of classifiers that can identify these signatures, it should be possible to eventually check every piece of luggage, cargo, and passengers that boards an aircraft. This device does not require the physical transfer of the explosive substance