Target spectra and maps

This section might need re-writing after constrained optimization is working.

The part of the program shown in Fig. 25 allows one to get composition maps from already-known spectra, or alternatively from cluster spectra. If you know all spectra, you can get thickness maps using SVD by selecting Add target spectrum ``.csv'' or ``.xas'' to load any known target spectra as optical densities (``mu''). One can also use Add a list of target spectra from ``.lis'' file to load a list of spectra. The option Add ``flat'' target spectrum loads one ``flat,'' constant-at-all-energies spectrum for evaluating fit residuals. Add cluster spectra loads all cluster spectra from the current cluster analysis. Normalize spectrum to 1/mu normalizes displayed spectra. Remove target spectrum removes the displayed target spectrum. Move target spectrum up moves a displayed spectrum up in the list, while Move target spectrum down moves a displayed spectrum down in the list. The slider-bar in the lower-left corner enables one to change the number of significant components and component weights in displayed target spectra. There is also an option to Generate maps from raw spectra or spectra fitted with principal components (upper right corner).

Finally, Save plot and fit coeffs as ``.csv'' saves an Excel-readable .csv text file. Save plots as ``s.png'' saves portable network graphics .png images. Save maps as ``.png'' saves portable network graphics .png images of thickness maps. Save maps as NetCDF ``.nc'' saves NetCDF .nc images of the maps.

Figure 25: The Target widget used to calculate composition maps.