
 Dewarping omnidirectional images using 2D LookUp Tables

 1/10/99, 4/1/00, J. Gaspar (www.isr.ist.utl.pt/~jag)

----------------------------------------------------------

The demo in this directory shows dewarpings of omnidirectional images
both to panoramic and bird's eye views. The dewarpings are performed
using previously computed LookUp Tables.


To run the demo:
- unzip all files to a single directory
- run matlab, change the directory to the one where the files
  were unzipped and at the matlab prompt type "demo"
- press return to go through the 3 examples


Main matlab functions:

remap2d.m  - Remap one image using a 2D lookup table.
remap2di.m - Initialise one LUT from pgm-pI5-files.
pi5read.m  - Read a modified PGM-P5 (Portable Gray Map image) file from disk.


Notes:

The LookUp Tables (LUTs) were built using zero order interpolation
(to our applications this is ok but in some cases it is important to use
higher order interpolation).

Every LUT is represented as a single column vector when loaded
into matlab memory, while in the hard-disk it is saved as two pgm-pI5 images.

Pgm-pI5 is a modified pgm file to hold integer values in binary.
Original pgm standard saves integer values in text format which results
in very large files as compared to saving into a binary format.
