Import widget for reading in spectral image, spectrum by spectrum.
The image is 200x150, each pixel is a 512 channel spectrum, 2 bytes each. Byte order is Intel. 'Vector' means that the data is spectrum by spectrum rather than image by image.
New windows after importing the data cube.
What the cube looks like after
Image type, size, global minimum pixel value, global maximum pixel value, image size in kb, and name.
Global spectrum of cube, obtained by:
Spectrum corresponding to roughly square area outlined below.
where
above spectrum came from.
Entire spectrum selected with region slider.
Global image generaged with '*Cube* -> Image -> Sum from Region'.
Image corresponding to one peak of the spectrum.
Layers (multiple regions) partly chosen for global spectrum.
This slider has seven bands. The light blue bands on either end are blank areas of the slider. Each band can be moved by dragging its end or its middle. Adjacent bands shrink or grow appropriately.
Layer bands completely adjusted:
Seven layer data cube, resulting from above seven layers.
The next to the last slice, slice 5, is shown here.
Each slice shown in an individual window using
(images shown half size).
Score (or component) cube with images of the seven principal components.
Eigenvalues corresponding to score images.
Loadings or Eigenvectors. (The actual 'rotation' matrix used to transform the initial layer stack into the component image stack.
Score or component images (half size)
Reconstruction of the 7 layers with only three components:
Plot of spectrum for entire image
Very short rectangle in navigator to examine high ev tail of spectrum.
Magnified right half of spectrum
Cube reduced x4 with same information:
Summed spectrum for cube reduced x4