Fundamentals for SXT Pointing Software and Databases

					/ys/sxt/doc/pointing_suncen.txt
					(MDM 9-Mar-95)

There are two fundamental pieces to the registration and pointing
coordinate question.
	1. Coordiantes of the Center of the Sun.
	2. Coordinates of the Center of the Field of View.
Both values are saved and passed around in absolute CCD coordinates.

SXT_CEN returns the absolute CCD coordinates of the center of
the sun for any given time.  The results are completely independent
of PFI/FFI/ROI location/summation/SXT_PREP regisration or anything
else.  The only input required for SXT_CEN is the date and time (when
an index or roadmap is passed in, only the date and time are used).
There is one pair of values for any given time.  SXT_CEN returns 
item #1 shown above.

By default, GT_CORNER and GT_CENTER return the absolute CCD coordinate 
of the corner or center of the image(s) that you have read.  GT_CORNER
and GT_CENTER work on an index and there is a value per image.  It
returns item #2 shown above, and this applies to FFI and PFI images. 
These routines were required because of the "messy" way that on-chip
summation occurs.

There are switchable options on the GT_CENTER routine
to return the coordinates relative to the sun center.  /ANGLE will
return the number of arcseconds east/west and north/south of the
center of the FOV relative to the sun center. The /HELIO option also
takes the sun center coordinates into account when producing the
heliocentric coordinates of the center of the field of view.
The /FROM_SC option for GT_CORNER give the coordinates of the lower 
left pixel of the image(s) relative to the sun center coordinate 
in summed pixels (dependent on the resolution of the image).  

If you want the IDL coordinates of the center of the sun
within each of your individual images (ie: within the IDL 2-D 
array), then you would do:
  IDL> rel_cen = ( sxt_cen(index) - gt_corner(index) )
  IDL> rel_cen(0,*) = rel_cen(0,*) / 2.^gt_res(index)
  IDL> rel_cen(1,*) = rel_cen(1,*) / 2.^gt_res(index)
For example, let's say that you had a FOV which had the lower
left corner exactly at the center of the sun, and that GT_CORNER
returns (x0,y0) for that lower left corner. SXT_CEN
would return the exact same values x0,y0 (based only on the
TIME of the image), and so "rel_cen" would be (0,0) saying
that the coordinates of the center of the sun for that image 
is in "data(0,0)".  The /FROM_SC gives the same result, but with
the oposite sign.

Lets run through an example:

IDL> rd_xda,'/yd1/sfd_daily/sfd950130.0613',[9,10,11,12,13],index,data
IDL> print,gt_center(index)
      512.500      511.500
      514.500      511.500
      512.500      511.500
      512.500      511.500
      512.500      511.500
				The center for the second image is
				different from the rest because it is
				QR, and there is a "feature" with the
				on chip summing readout.

IDL> print,gt_center(index,/angle)
     -57.2020     -172.278
     -52.6170     -187.886
     -58.7685     -185.778
     -59.4930     -183.928
     -60.1687     -180.851
				Since these are un-registered, the position
				relative to the suncenter wanders.
IDL> print,gt_corner(index,/from_sc)
% Compiled module: GT_SXT_CEN.
GT_SXT_CEN: An index with a .HIS was passed in, but no
GT_SXT_CEN:     values in .SUN_CENTER so SXT_CEN is being called
     -267.870     -290.375
     -133.250     -146.538
     -268.245     -293.120
     -268.385     -292.740
     -268.510     -292.110
				The /FROM_SC option for GT_CORNER give
				the coordinates of the lower left pixel
				of the image(s) relative to the sun center
				coordinate in summed pixels (dependent on
				the resolution of the image).  The sun
				center coordinates are (0,0) and the 
				value returned is relative to that, so for
				an FFI, the lower left corner is in the
				"negative" direction from sun center.

				Now lets register the images and see
				what the results look like.
IDL> sxt_prep,index,data,index2,data2,/reg,outres=2
IDL> print,gt_center(index2,/angle)
     -58.6552     -182.003
     -58.6554     -182.003
     -58.6554     -182.003
     -58.6554     -182.003
     -58.6554     -182.003
				Since the image has been registered,
				the coordinates of the center of the field
				of view are all equidistant from the center
				of the sun.
IDL> print,gt_corner(index2,/from_sc)
     -133.751     -145.942
     -133.751     -145.942
     -133.751     -145.942
     -133.751     -145.942
     -133.751     -145.942


A ton of the problems that people tend to have is resolving what is
relative to what.  All coordinates that we use (SXT_CEN, GT_CORNER, 
and GT_CENTER) are in absolute pixel coordinates on the fixed CCD.  
When we register an image, then GT_CORNER-SXT_CEN all give the same 
results, because those numbers are relative to sun center.  The other 
thing to remember is that these absolute coordinate from SXT_CEN and 
GT_CORNER have nothing to do with the IDL subscripts in the "data" 
variable.  That is probably where most people have problems.  If you 
want the sun center coordinates relative to the lower left corner 
pixel (ie: "data(0,0)" then you would do SXT_CEN - GT_CORNER, so that 
you have pixels RELATIVE TO THE DATA ARRAY, where GT_CORNER - SXT_CEN 
is RELATIVE TO SUN CENTER.It's all a matter of what you want relative to
what.  We save everything in the same coordinate system and units so that
we can go to and from all of these other coordinate systems.