Spectral Imaging: Methods, Design, and Applications - Biomedical Optical Imaging Technologies: Design and Applications - page 119

Fig. 4.2 Description of a
spectral image data set. Each
point in the cube represents a
single number, and the
spectral image is described as
I.x; y; /. It can be viewed
either as an image I.x; y/ at
each wavelength or as a
spectrum I./ at every pixel
(x; y)
in the development of many different spectral imaging techniques differing from
one to another by the method in which the spectral image is collected. Each of the
methods has its advantages and limitations which depend on the character of the
application for which it is being used. In that sense, coming to choose a spectral
imaging system is rather a matter of tailoring the most adequate spectral imaging
method to the type of application.
Spectral imaging methods can be divided into the following methods:
1. Methods that capture a full spectral image at very few spectral bands simulta-
neously. The disadvantage is that such methods compromise on the number of
points in the spectrum, the field of view, and spatial or spectral resolution. Such a
measurement can be represented as I.x; y; /,where is measured at only few
wavelengths, typically of the order of 3 or 4.
2. Wavelength-scan methods that measure the images one wavelength at a time.
Here, each measurement i measures I.x; y; i / at a single wavelength i ,and
the measurement is repeated n times. To span a wavelength range, typically
30-40 different acquisitions should suffice.
3. Spatial-scan methods that measure the whole spectrum of a portion of the image
at a time and scan the image (e.g., line by line). In this case the measurement
is I.x; y i ;/,wherey i represents a single point along the y axis, and the
measurement repeats n times to cover the whole required range along y.
4. Time-scan methods that measure a set of images where each one of them is
a superposition of spectral or spatial image information. The actual spectral
information is calculated at the end of the acquisition by a mathematical trans-
formation such as Fourier transform. Here, the measurement is of I.x; y; f i .//,
where f i ./ is a function of the spectrum at every pixel and in the end it is
transformed to give I.x; y; /. The method that is based on Fourier spectroscopy
belongs to this category. In this method, the light coming from every pixel passes
an interferometer that changes its intensity according to the time delay created in
between the two arms of the interferometer. This information is then transformed
to the spectrum, see Sect. 4.4.4 .
5. Spectral images usually measure the emission (or transmission) from the sam-
ples, but excitation spectra can also be obtained by placing the system on the
excitation optical port. For transmission measurements, such a method provides
similar information, but it is different for fluorescence. Dickinson et al. [ 8 ]have