Last edited on 2009-01-29 22:26:22 by stolfi

Processing the IAB-2002 fragment pictures
Camera calibration and image set integration

=== TO REDO ===
  
IMAGE ALIGNMENT

  Images from the "stdu" batch were divided into sub-batches, the
  "light-groups", so that within each group the only difference was in
  the lighting (L = "a", "b", or "c").

  Within batches stdu/000 through stdu/012, each light group was
  aligned by picking refrerence points with great care and applying
  translations, small rotations, slight scaling, and trimming to a
  common size. The scripts align-all-images and align-images were used
  for that purpose.
  
  The results are work/align/%B-%S%E%L-L.png . The common trimmed size for each light
  group is work/align/%B-%S%E.size
  
COMBINING IMAGES WITH DIFFERENT LIGHTING 
  
  The aligned images in each light-group were normalized and averaged to produce
  work/avglight/%B-%S%Em-L.png

======================================================================
%% NEVER DONE   
    

IMAGE ALIGNMENT
   

  (Probably do a second stage, after determining the exact position 
  of the grid, comparing the computed model against the image.)