$Id: iros.html,v 1.1 1997/01/20 13:42:27 jaaps Exp jaaps $


iros, extract X-ray source list form detector image


iros (Iterative Removal Of Sources) reconstruct celestial image for a detector-image.


The program can be started with no arguments at all, in which case the parameter values are taken from the parameterfile.
Alternatively, values can be passed via the commandline, the required items as command arguments, the optional ones as valued switches.

The last form (iros -q) creates the parameterfile in $PFILES.

The table lists all the parameters used by iros, the second column lists the paramater key names used in the parameterfile, the first column gives the corresponding switch letter.

iros parameters
expfile stringi input di, detector image file.
resfile stringo output, reduced sky and det image, sourcelist .res
aislfile stringi initial list, accepted as genuine sources, catalog format.
Axysilst stringi extra initial sources, listing optional name, position and spectral index: name,X,Y,si.... , name is optional, if present has to start with a letter.
battfile stringi attitude data file, overrides default path/name
ccmode inti correlation mode, correlate
ddecorum stringi set (1), to restore subtracted Xray source(s) pixels in the output skyimage.
echnlset stringi for TSPZ(TeleSPaZio)-format detectorimage input only, set the channel (energy) selectionwindow. the header of these detector images contains no such information, it is needed in the psf computation..
fofits inti by default, if inputimage is in TSPZ format, output will be too, set to force force FITS output.
hatmhght reali earth atmosphere thickness (km), used in source visibility (earth occultation) determination.
jskyrsltn ipairi skypixels can be different sources if they are separated by at least this many pixels.
kcatfile stringi overrides default X-ray source catalog (environment).
miros inti
1 suppress use of X-ray source catalogs.
2 interactive operation, subtract detected sources one by one, prompting the user for position, flux and pointspread function modification.
4 use all detected sources in second and next iterations, whether this source could be identified or not.
8 accept any pixelconfiguartion as valid source.
16 suppres use of source list found as (FITS)extension in input detector image (usually simulated data).
32 treat detector as having no pointspread function, all pixels from a reconstructed x-ray source in the celestial image are treated separately.
64 subtract only sources/pixels when found in catalog.
nniros inti absolute limit on number of iros iterations, default is 15, but if niros is not assigned iros will also stop at an increase in the detecor rms (root-mean-square).
pp_filtr areai include section, x0-x1:y0-y1[,x2-x3,y2-y3 .... ], from input detectorimage.
pxp_filtr areai As above, but exclude section.
rradius reali maximum radial distance (pixels) between a candidate- and a catalog X-ray source for identification.
ssrclim inti limit number of sources considered to subtract.
tthrshld reali stop criterium, threshold, stop search for more sources, if the combined detected sources account for more than a fraction 'r' of the total counts on the detector.
Vvoigt rpairi voigt function parameters (pointspreadfunction model) , DO NOT USE - not calibrated.
xpointing stringi override any pointing from input (.exp) file, format: ra,dec.na, in decimal degrees.
yinlcrcn inti set to 1 to use integral detector non-linearity data.
Zgssnfctr rpairi general multiplication factors, detector gaussian response, better not use.
-xchisq inti NOT used in this releasedet. fit control,
Iinsfile stringi overrides the IDF name in expfile.

p, L, i StageII global switches


iros, 'Iterative Removal of Sources', is a method to retrieve the image of the sky from an detector image, produced by an exposure of the Wide Field Camera to this sky. The method can be depicted like: At the end of this process the positions of, presumably all, say n sources in the field of view are determined, and we can write the following expression for the expected value of all N (6802) detector pixels d:

d[1]= S1[1]*f1 + S2[1]*f2 + ..... Sn[1]*fn + Sb[1]*fb.
d[2]= S1[2]*f1 + S2[2]*f2 + ..... Sn[2]*fn + Sb[2]*fb.

d[k]= S1[k]*f1 + S2[k]*f2 + ..... Sn[k]*fn + Sb[k]*fb. d[N]= S1[N]*f1 + S2[N]*f2 + ..... Sn[N]*fn + Sb[N]*fb.


The computation of the Si[k] values requires a simulation of the Wide Field Camera, where we consider: This is exactly the same procedure as used in simulate step from the iros procedure above, only here we must preserve the detector image for each source. After computation of all Si[k], we have an overdetermined system of equations: the least-squares solution of which is So finally, we are left with


We have an exposed detector image w817w1.exp, the result of an cdi run. The command (assuming valid environment variables are set)
iros -p2 w817w1 w817w1
produces the following output
 iros[99.99]                               Fri Jul  4 13:12:53 1997  page    1
[2]src_readcatalog: "/ruurq2/st2/lib/catalogs/XRAYSOURCESCAT", 7 potential sources in FOV
   "OP817(wfc1)"   Fri Aug 23 11:42:00 1996 (UT)
   "ENP + WFC gal.ce"  @ [05h41m05.304s +29d04m08.195s (na: 178.754)]
   iros: plane 1 (channels 2 - 18)
[1]iros_loop<00>: detector sum=   497894.0 (100.0%).  rms= 1.960604e+00
[1]src_showsl: iros_reduce: subtracting  (1).
 + seqn        name       ra     dec      -x-     -y-    -flux- signif
 + < 1>     X0531+219   83.614  22.078   86.56   17.14    0.472 1068.2 ([341,271]3x2)
[1]iros_loop<01>: detector sum=   246505.4 ( 49.5%).  rms= 1.275229e+00 ( 1 src subtracted)
[1]src_showsl: iros_reduce: subtracting  (1).
 + seqn        name       ra     dec      -x-     -y-    -flux- signif
 + < 1>     X0531+219   83.616  22.077   86.58   17.12    0.527 1220.6 ([341,271]3x2)
[1]iros_loop<02>: detector sum=   216883.4 ( 43.6%).  rms= 1.240710e+00 ( 1 src subtracted)
[1]src_showsl: iros_reduce: subtracting  (1).
 + seqn        name       ra     dec      -x-     -y-    -flux- signif
 + < 1>     X0531+219   83.616  22.076   86.59   17.11    0.532 1252.0 ([341,271]3x2)
[1]iros_loop<03>: detector sum=   214217.8 ( 43.0%).  rms= 1.238627e+00 ( 1 src subtracted)
[1]src_showsl: iros_reduce: subtracting  (1).
 + seqn        name       ra     dec      -x-     -y-    -flux- signif
 + < 1>     X0531+219   83.616  22.076   86.59   17.11    0.532 1268.8 ([341,271]3x2)
[1]iros_loop<04>: detector sum=   214581.9 ( 43.1%).  rms= 1.239049e+00 ( 1 src subtracted)
[2]iros_solve final chisquare: 1.162030
   iros_loop: 4 iterations, final detector sum= -5003.5, -1.00%.  rms= 1.118583e+00
[1]src_showsl: iros: final sourcelist (2).
 + seqn        name       ra     dec      -x-     -y-    -flux- signif
 + < 1>     X0531+219   83.616  22.076   86.59   17.11    0.506 1252.0 ([341,271]3x2)
 + < 2>   :background:  85.272  29.069    0.00    0.00    1.298    0.0 ([  0,  0]0x0)
[1]evt_open: "/ruurq2/817/WOP817_0000w1att1.fits"[ATTITUDE], 1 of 1.
 +          time  50318.48730 - 50318.56047 (MJD)  [  6322.00 sec.]
[2]iros_effetime: X0531+219 : 2778.38 sec. (51.9%)
[2]iros_effetime: :background:: 2721.38 sec. (50.8%)
and produces a file w817w1.res. The ftool fstruct run on w817w1.res produces
  No. Type     EXTNAME      BITPIX Dimensions(columns)      PCOUNT  GCOUNT
   0  PRIMARY                -32     510 510 2                   0    1
   1  IMAGE    rdi           -32     680 680                     0    1
   2  BINTABLE WFC_Sourcelis   8     1384(19) 2                  0    1
the first extension, rdi, is the residual detector image. the rdi is the difference between the measured, input, detector image and a simulated image of the detector, using the detected sources and the background from the final iros run.

the PRIMARY image in the .res file is the so-called residual skyimage, the correlation of the reduced detector image and the mask, any sources showing up in this image were not identifyable, as there is no entry for these sources in the catalog.

the binary table extension WFC_Sourcelist (fstruct list only 13 char's in the name) harbours the list of all detected sources. the table has the following structure and a row for every detected source plus one for the background:
iros columns
ra 1Dright ascension, degrees. (J2000,FK5)
dec 1Ddeclination, degrees.
xtm 1Dexposure time, seconds, corrected for mean ltf and earth occultation.
cname 32Asax-wfc canonical source name, X0000-000
tname 32Atrivial source name.
flux 32Esource flux estimate, cts/sec.cm2.
dflux 32Eerror in source flux, cts/sec.cm2, 1 sigma level established in source - pointspread model fit.
x 32Esource offset from center, in skypixels.
dx 32Eerror in above offset, in skypixels, 1 sigma level established in source - pointspread model fit.
y 32Esource offset from center, in skypixels.
dy 32Eerror in above offset, in skypixels, 1 sigma level established in source - pointspread model fit.
sigma 32Epure poisson error in source flux, square root of total number of events used in flux computation.
signif 32Esource significance, flux divided by above defined sigma.
chisq 32Esource - pointspread model fit, final chi square.
F 32E"goodness of fit", chisq reduction in comparison to flat model.
type 1Jspectral type, copied from input catalog.
nh 1EH column density, copied from input catalog.
index 1Espectral index, copied from input catalog.
spare 1J

though it does look promising, the flux and pixel position related columns have 32 entries, suggesting full energy resolution in all 32 channels, this is NOT automatically the case. In the current iros revision, sources fluxes are determined in the same energy bands that were selected in the input .exp file, there is no extrapolation to other channels.
Only the first m positions in the flux .. F columns are filled, when the input file has m detector images. So. if you want full energy resolution, you must first use cdi to produce a .exp file with 31 detector images, one for each channel. (channel 0 does not occur in the WFC event data.)
Note: The amount of memory needed in cdi is then quite large, for the images only ca. 31*6802*4 = 60 Mb, greater problems loom in iros, this program has, in this version, still some memory leaks, it is likely to crash eventually. In all, its better to break up the problem in a few smaller steps.

  • -linearisation-
  • -binning- as defined in the instrument-file(FITS)



    Back to: