Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.adass.org/adass/proceedings/adass94/vanstoned.ps Äàòà èçìåíåíèÿ: Tue Jun 13 20:55:43 1995 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 02:27:14 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï

Astronomical Data Analysis Software and Systems IV
ASP Conference Series, Vol. 77, 1995
R. A. Shaw, H. E. Payne, and J. J. E. Hayes, eds.
Recreating Einstein Level One Processing Exposure
Masks and Background Maps in IRAF
D. Van Stone, M. Garcia, J. McDowell
Center for Astrophysics, 60 Garden St., Cambridge, MA 02138
Abstract. This paper describes the main algorithms used by the Ein­
stein Level One processing to create the exposure masks and the back­
ground maps for Einstein IPC images, and how these algorithms were
recreated in the IRAF environment.
1. Introduction
Our goal was to recreate the algorithms used by the Level One Processing to
create exposure masks and background maps for Einstein Image Proportional
Counter (IPC) data (cf. sections 2.5 and 2.7 of Harnden et al. 1984). Although
these images have not been archived on CD­ROMs, the aspect and original
background image information have been. From this information, one can ei­
ther recreate the original background images, or recalculate and create exposure
masks and background maps from within IRAF. Special consideration is given
to extend the background map algorithm to work on the Einstein unscreened
IPC images 1 .
2. Exposure Masks
The aspect information for an image contains a timeline of the aspect solution
for the duration of the observation. The exposure mask is simply the sum of
multiple images, where each image is the result of applying an aspect solution
to the IPC geometry, weighted by that aspect's duration. This algorithm does
not attempt to correct for vignetting.
2.1. Exposure Mask Algorithm
For each set of aspect data and duration, an exposure mask is created by look­
ing at each pixel in the final exposure mask, and finding the pixel in detector
coordinates by de­applying aspect, and seeing that if the detector coordinates
lie within the IPC geometry 2 , add duration of aspect data to the image pixel.
1 The unscreened images were published on CD­ROMs in March 1994 (McDowell et al. 1994).
2
The IPC geometry is defined as all pixels 287 Ÿ x Ÿ 737 and 288 Ÿ y Ÿ 738 which are at least
15 pixels away from the rib centers of x = 359:6, x = 656:8, y = 376:6 and y = 673:8. All
values are in PROS coordinates, the system used on the CD­ROM archive.
1

2
Figure 1. The Einstein unscreened IPC image for sequence I5803,
along with its exposure mask created by the EINTOOLS task
exp make.
3. Background Maps
The main idea behind the background map algorithm is that the background
for an individual image can be modeled as some combination of blank field
(``deep survey'') images and flat field (``bright Earth'') images. The reason that
the blank field image by itself is not a sufficient model is that each image (in­
cluding the blank field) has a variable amount of particle induced and/or stray
light background. Our algorithm accounts for this by including some (possibly
negative!) amount of flat field data in the model background.
We assume that the blank field rate is constant over time, and thus the
normalization we use for the deep survey image is based solely on the duration
of the exposure. The normalization for the bright Earth image is set so that the
total number of background counts (excluding sources) in the model is the same
as those in the image, as determined by the following relation:
image counts = source counts + (deep survey counts + bright Earth counts)
Note that the background map algorithm does not explicitly take into account
internal or particle induced background.
3.1. Background Map Algorithm
The algorithm used to create a background map is as follows: (1) calculate
counts in the image due to sources 3 (SRC CNTS); (2) calculate counts in deep
survey and bright Earth maps 4 (DS CNTS & BE CNTS); then (3) for each set
of aspect data and duration:
3
See section 3.2.
4
The counts are calculated using the bright edge region. See section 3.3.

3
1. Find number of photons in image which arrived during this set of aspect
data (IM CNTS)
2. Calculate deep survey factor (DSFAC):
DSFAC = (aspect duration)
(deep survey livetime)
3. Calculate bright Earth factor (BEFAC):
(BEFAC) \Theta (BE CNTS) = (IM CNTS) \Gamma (DSFAC) \Theta (DS CNTS) \Gamma
(aspect duration)
(image livetime)
(SRC CNTS)
4. Apply BEFAC and DSFAC weights to bright Earth map and deep survey
maps and sum these two images
5. Apply aspect to the summed image and add to background map
3.2. Source Counts Algorithm
The algorithm used to calculate the number of counts for each source in the
image is; (1) calculate total counts in the source circle and in the background
annulus 5 (TOT CNTS & BG CNTS); (2) calculate the area in the source circle
and in the background annulus 6 (SRC AREA & BG AREA), and (3) compute
the counts attributable to this source by the relation:
(SRC CNTS) = CCC \Theta
`
(TOT CNTS) \Gamma (BG CNTS) (SRC AREA)
(BG AREA)
'
where CCC, the circle composite correction, is defined as the product
CCC = (circle mirror scattering corr.)(circle point response corr.):
The user can generate a list of sources by using the PROS task LDETECT.
3.3. Masking Detector Hotspots
Our background map algorithm works best if the internal (particle induced)
background is small, at a constant rate, and uniform over the field. In order to
make this true, we must exclude the regions of very high internal background
on the edges of the detector from the calculation of the normalization. This
exclusion is performed with the bright edge filter on QPOE files and the bright
edge region on the deep survey and bright Earth images 7 .
5
These counts are calculated using the bright edge filter. See section 3.3.
6
All areas are calculated using the exposure mask.
7
The bright edge filter and bright edge region describe identical areas. They differ because one is
for filtering detector coordinates within QPOE files whereas the other is a region on an image.

4
Figure 2. The figure to the left is the broad band background map for
the Einstein unscreened IPC image I5803, created by the EINTOOLS
task bkfac make. The figure to the right is the original Level One Pro­
cessing background map created by the EINTOOLS task be ds rotate.
4. Eintools
The new PROS package EINTOOLS contains tasks for creating exposure masks
and background maps for any Einstein IPC QPOE file. One can also create the
original Level One Processing background maps using the bright Earth and deep
survey factors from the original processing. The tasks allow the user to screen
on any time filter, produce background maps for any (possibly nonstandard) PI
range, and create exposure masks of any specified resolution.
This package will be available in the next major release of PROS. It is our
hope that some of the data structures created for this package (such as a generic
time­resolved aspect table) will be used in future programs generating exposure
masks or background maps.
Acknowledgments. The PROS project is partially supported by NASA
contracts NAS5­30934 (RSDC) and NAS8­30751 (Einstein).
References
Harnden, Jr., F. R., Fabricant, D. G., Harris, D. E., & Schwarz, J. 1984, Scien­
tific Specification of the Data Analysis System for the Einstein Observa­
tory (HEAO­2) Imaging Proportional Counter, SAO Special Report 393,
(Cambridge, Smithsonian Astrophysical Observatory)
McDowell, J., Plummer, D., Prestwich, A., Manning, K., Van Stone, D., & Gar­
cia, M. 1994, The Einstein Observatory IPC Unscreened Data Archive,
CD­ROM Volumes 0--18, (Cambridge, Smithsonian Astrophysical Obser­
vatory)