iraf.net

Hi, All

I am using geomap and geotran to align two images. The xshift and yshift which obtained by geomap is less than 1 pixel. But when I transform one of images to match another, there is a shift which is much larger than 1pixel in the resultant image.

The result seems to be independent of the geomap parameters. The images are DECam datas. Did the result caused by the data type?

Geomap result:

# Sat 22:58:53 20-Feb-2016
begin a
xrefmean 1900.46533203125
yrefmean 1066.5546875
xmean 1900.409912109375
ymean 1066.453735351562
geometry general
function polynomial
xshift -0.1340466
yshift 0.1107178
xmag 1.000012
ymag 0.9998681
xrotation 0.002140492
yrotation 0.002997969
xrms 0.2416764
yrms 0.4075866
surface1 11
3. 3.
2. 2.
2. 2.
0. 0.
0. 0.
5000. 5000.
0. 0.
5000. 5000.
-0.1340466 0.1107178
1.000012 -3.735907E-5
5.231752E-5 0.9998681
surface2 23
3. 3.
5. 5.
5. 5.
2. 2.
0. 0.
5000. 5000.
0. 0.
5000. 5000.
2.158698 1.701186
-0.003266457 -0.002010282
1.856530E-6 7.470235E-7
-4.299098E-10 -9.053586E-11
3.926483E-14 -6.083670E-16
-0.005131495 -0.00601719
3.443224E-6 4.423120E-6
-9.756651E-10 -1.112400E-9
6.401939E-14 9.199640E-14
4.557282E-6 7.245617E-6
-1.682136E-9 -2.635244E-9
2.342344E-13 3.192638E-13
-1.351058E-9 -3.718675E-9
2.871175E-13 4.325086E-13
3.394932E-14 7.127380E-13

Geotran parameter:

input = "aaa.imh" Input data ### I have tried to use fits and imh files as the input of geotran. But it did not work.
output = "bbb.fits" Output data
database = "database" Name of GEOMAP database file
transforms = "a" Names of coordinate transforms in database file
(geometry = "geometric") Transformation type (linear,geometric)
(xin = INDEF) X origin of input frame in pixels
(yin = INDEF) Y origin of input frame in pixels
(xshift = INDEF) X origin shift in pixels
(yshift = INDEF) Y origin shift in pixels
(xout = INDEF) X origin of output frame in reference units
(yout = INDEF) Y origin of output frame in reference units
(xmag = INDEF) X scale of input picture in pixels per reference unit
(ymag = INDEF) Y scale of input picture in pixels per reference unit
(xrotation = INDEF) X axis rotation in degrees
(yrotation = INDEF) Y axis rotation in degrees
(xmin = INDEF) Minimum reference x value of output picture
(xmax = INDEF) Maximum reference x value of output picture
(ymin = INDEF) Minimum reference y value of output picture
(ymax = INDEF) Maximum reference y value of output picture
(xscale = 1.0) X scale of output picture in reference units per pixel
(yscale = 1.0) Y scale of output picture in reference units per pixel
(ncols = INDEF) Number of columns in the output picture
(nlines = INDEF) Number of lines in the output picture
(xsample = 1.0) Coordinate surface sampling interval in x
(ysample = 1.0) Coordinate surface sampling interval in y
(interpolant = "linear") Interpolant
(boundary = "constant") Boundary extension (nearest,constant,reflect,wrap)
(constant = 0.0) Constant boundary extension
(fluxconserve = yes) Preserve image flux?
(nxblock = 4096) X dimension of working block size in pixels
(nyblock = 4096) Y dimension of working block size in pixels
(verbose = yes) Print messages about the progress of the task
(mode = "ql")

Any help would be greatly appreciate.

Verda

Hello Verda,

GEOTRAN can be tricky to use because if you just use defaults it can make the image larger. This would mean that in terms of pixel coordinates there can be an apparent mismatch even though in world coordinates and any distortions are properly accounted for. I can't deduce what is needed just from the information you sent, though what you did send is important. I can look at this if you either send me the images (valdes@noao.edu), put them somewhere I can get them, or just send a long header listing (imhead aa l+) from which I can construct a dummy image.

This would definitely not be related to image formats used.

Yours,
Frank Valdes

Dear Frank,

Thank you so much for your help. I have solved the problem. I found that the keyword of header 'CRPIX' has changed at the same time. That is why a larger shift happened in the resultant image. Thank you very much again!!!

Sincerely yours,
Verda