Hyperspectral Applications
SOC's hyperspectral technologies have been applied in such diverse fields as environmental monitoring, camouflage detection, land mine detection, counterfeit currency and cannibas detection.
Pollutant Detection in Drainage System
SOC worked with local officials in Pennsylvania to determine the presence of a suspect pollutant in a drainage system. The test data was evaluated to determine if it was possible to construct a detection system for locating pollutants in and along waterways.
RGB Image
RGB Image Processed with Spectral Filter
Contaminated Water Indicated by
processing Hyperspectral ImageryWith Spectrally Matched Filter
Camouflaged Target Detection
An example of hyperspectral detection is shown at right where a scale model M1A1 tank has been painted with a carefully tailored pigment paint to match the background in which it was placed. As noted in the spectral data, there is little difference between the two and the match is extremely close. The hidden tank model cannot be seen in the picture on the left. However, applying spectral correlation techniques to the image immediately locates the hidden object (right picture frame).
Spectral detection of a camouflaged tank model.
Land Mine Detection
Another example of spectral correlation is shown at right. Each material in the scene exhibits its particular spectral response as noted in the spectral curves. Each pixel in the scene can be related to that material via its spectral response. In this instance, the various forms of vegetation in the scene, can be identified and subtracted spectrally form the scene through software algorithms to reveal a partially concealed man-made object (landmine). This process illustrates the methodology for separating "target" items from their backgrounds. The minutely different spectra between target and background show up as "bright" pixels in image as in the case with the model tank.
Spectral correlation makes it possible
to identify 4 mines in complex scene
Counterfeit Currency Detection
The US
Treasury Department has incorporated a number of features in
its new currency issues for security measures against advanced
technologies in counterfeiting. These features serve as a
basis of priori knowledge of the image and the spectral
response in each image pixel. They include:
- Image: Detail, Size, Location, Alignment
- Numerics: Size, Location, Alignment
- Paper Type
- Ink Selection
- Watermark
- Characters: Size, Location, Bidirectional Optical Properties
- Other: Issuing Agency Features
Use of ink as a detectable feature in counterfeit currency
To illustrate the
effectiveness of the Hyperspectral Detection System against
counterfeiting, copies of currencies were obtained through
legal methods and compared spectrally to real currency of the
same denomination. Detection principles are illustrated in
the figures at right using only one of the features in the list above,
that being the comparison of real ink and counterfeit ink as
viewed by a hyperspectral sensor. As noted, there is a large
difference in the spectral response of the two inks at the
longer wavelengths of the visible band, and the counterfeit
bills are easily detected.
Comparison of ink spectral responses, real and counterfeit.
Real and counterfeit currency with no spectral
discrimination.
Hyperspectral Imaging reveals differences between real
and copied currency.
Marijuana / Cannabis Detection
Plants also have discriminating spectral features which can lead to their detection in a background of other plants as shown at right. In this illustration, the ability of the SOC imagery technology is evident in detecting the presence of an illegal cannabis-type plant. This demonstration was recently organized and carried out by local law enforcement officers using SOC hyperspectral camera technology.
Detection of a Cannabis-Type plant
in naturally occurring vegetation.
Surface Optics
Corporation 11555 Rancho Bernardo
Rd.San Diego, CA
92127
Phone:
858.675.7404 Fax: 858.675.2028
copyright 1996-2001 by Surface Optics Corp.
All Rights
Reserved