27 Aug 1993

This file gives a general overview of how the three programs:
i.points3      - modified i.points program
i.rectify3     - modified i.rectify program
i.tape.tm3     - modified i.tape.tm program.
were designed to work.


-------------------------------------------------------------------------------
THE BASICS

  The programs utilize the GRASS Database hierachy as usual.  

  $GISBASE
     |
  $LOCATION
     |
  $MAPSET 

  Recall that LOCATION can be one of several projection types.
  A - xy 
  B - UTM
  C - State Plane
  D - lat/lon
  E - Other

  Raster, vector, and site data are stored in sub-directories under 
  a users MAPSET. 

  Also under a users MAPSET, a sub-directory known as an "imagery group"
  or simply "group" can be created with the "i.group" program.  Imagery
  groups are used from GRASS image processing programs.  

  i.points3, i.rectify3, and i.tape.tm3 all use imagery group files.


-------------------------------------------------------------------------------
I.TAPE.TM3
  i.tape.tm3 is a data extraction program for reading LANDSAT Thematic Mapper
data sets.  The data set may be contained on tape or files.

  i.tape.tm3 works only for xy LOCATIONS.  

  i.tape.tm3 will extract the imagery from the data set without performing any
modifications (stretching, rotating, ...).  If you extract a 1000x1000 piece
on imagery from the data set, you will get a 1000x1000 raster file.


  A single band of LANDSAT TM data is stored on the original data set as:

  row 1 -------------------------------------------------------
        |            xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx   |
        |           xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    |
        |           xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    |
        |          xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx     |
        |         xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx      |
        |         xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx      |
        |        xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx       |
        |       xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx        |
        |      xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx         |
        |     xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx          |
        |    xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx           |
        |   xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx            |
        |  xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx             |
  row N -------------------------------------------------------

       col 1                                                col M

  where the "x" represent data and everything else is padding.
  When the full imagery is extracted, it will be placed into 
  a raster file with the boundaries as shown.

                           North = N + 0.5
        -------------------------------------------------------
        |            xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx   |
        |           xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    |
        |           xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    |
        |          xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx     |
        |         xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx      |
        |         xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx      |
  East  |        xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx       | West
  0.5   |       xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx        | M + 0.5
        |      xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx         |
        |     xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx          |
        |    xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx           |
        |   xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx            |
        |  xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx             |
        --------------------------------------------------------
                           South = 0.5



  To extract only a piece of the original data you can specify
  the starting and ending rows and columns:
  For example:
      i.tape.tm3 .... rows=1,200  cols=500,1000 ...
  would extract the areas indicated by "x"  or "o" below:

  row 1 -------------------------------------------------------
        |         ooooxxxxxxxxxxxxxxxxxx                       |
        |         oooxxxxxxxxxxxxxxxxxxx                       |
        |         ooxxxxxxxxxxxxxxxxxxxx                       |
        |         oxxxxxxxxxxxxxxxxxxxxx                       |
row 100 |         xxxxxxxxxxxxxxxxxxxxxx                       |
        |                                                      |
        |                                                      |
        |                                                      |
        |                                                      |
        |                                                      |
        |                                                      |
        |                                                      |
        |                                                      |
  row N -------------------------------------------------------
                  |                     |
       col 1    col 500             col 1000               col M

  where the "x" represents original data and "o" is original padding.

  This example will result in  a raster file with the boundaries as 
  shown below.

         North = (N+0.5) - start_row + 1 
               = (N+0.5)               
     
                 ooooxxxxxxxxxxxxxxxxxx                       
                 oooxxxxxxxxxxxxxxxxxxx                       
 East =          ooxxxxxxxxxxxxxxxxxxxx  West =                      
 (0.5) +         oxxxxxxxxxxxxxxxxxxxxx  (0.5) +                     
 start_col       xxxxxxxxxxxxxxxxxxxxxx  end_col                     

         South  = (N+0.5) - end_row + 1
                = (N+0.5) - 99


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

LANDSAT TM data comes on a tape with the following structure

   
#1    VOLUME HEADER
#2    Band 1 header
#3    Band 1 data
#4    Band 1 trailer
#5    Band 2 header
#6    Band 2 data
#7    Band 2 trailer
   ............
#     Band 7 header
#     Band 7 data
#     Band 7 trailer

#     NULL FILE (EOF)
