Althoughhyperspectral cameras are versatile instruments with applications in a wide variety of monitoring and analysis applications, their cost and complexity have limited their use.
Now, researchers atAberystwyth University (Aberystwyth, UK) have devised a simpler and more economical hyperspectral camera that would be suitable for applications ranging from agricultural monitoring to exploring the surface of Mars.
Hyperspectral cameras capture data sets known as hypercubes -- stacks of 2-D image data at different wavelengths. Since detectors can only capture two dimensions of data at one time, the third dimension of the hypercube must be obtained by scanning.
The two established methods of capturing a hypercube are wavelength scanning in which 2-D images are captured at each wavelength in turn (see figure, a) and image scanning in which one image dimension and the wavelength dimension are captured simultaneously and the second image dimension is scanned (see figure, b).
For wavelength scanning, electro–optical devices such as liquid crystal and acousto-optic tunable filters (LCTF and AOTF) are incorporated into the camera to select the wavelength for each monochromatic image. For image scanning, an imaging spectrograph disperses the spectrum across one dimension of the detector with an image of the spectrograph slit formed across the other dimension. The second dimension of the image is obtained by scanning the camera.
The proposed method for the new hyperspectral imager developed at Aberystwyth University employs a hybrid of the image and wavelength scanning methods (see figure, c). An interference filter mounted at an angle in front of a monochrome camera disperses a region of the spectrum across the field Of view (FOV) so that an image is obtained with different regions captured at different wavelengths.
By scanning the camera across the FOV, all regions are imaged at all wavelengths within the range. An extended wavelength range could be also obtained by employing a set of filters mounted in a tilted filter wheel.
Aberystwyth researchers Matt Gunn, Dave Langstaff, and Dave Barnes claim that the method has a number of advantages over the existing methods including higher light throughput, no polarization sensitivity, low power consumption, lighter weight, a high degree of robustness and reliability, and relatively low cost (see details here).
-- By Dave Wilson, Senior Editor,Vision Systems Design
