Spatial Analysis

New command CREATE/WOBBLED_MASK (Jun 2, 1995)

Purpose of CREATE/WOBBLED_MASK

CREATE/WOBBLED_MASK produces a 0,1 mask to mask out areas obscured by the PSPC ribs. This has to be done to avoid unreliable results when running source detection. The source detection doesn't take into account the obscuration by the PSPC ribs, so applying source detection to a source partially obscured by a rib leads to wrong results.

The shadow of the PSPC rib structure is smeared out along the wobbling direction. Therefore simple rotating and aligning an existing map structure without taking this wobble into account (as in ADJUST/MASK) yields a poor match for rib structures perpendicular to the wobbling direction. CREATE/WOBBLED_MASK corrects for this. It is recommended that you replace ADJUST/MASK with CREATE/WOBBLED_MASK when analyzing your PSPC data.

CREATE/WOBBLED_MASK doesn't work for HRI data. However, here is no need for that.

How CREATE/WOBBLED_MASK works

In the EXSAS calibration area there are two new 0,1 maps of the PSPC rib structure, monew_b.bdf and monew_c.bdf for PSPC B and C, respectively. These maps were constructed from the actual PSPC instrument maps (courtesy W. Pietsch).

CREATE/WOBBLED_MASK takes one of those, wobbles and sums them according to the contents of the attitude table. This is done internally by calling command CREATE/EXPOSURE with a parameter file crewob.par, which is constructed by CREATE/WOBBLED_MASK. Time selections and image binsizes are taken from the descriptors of the reference image. The resulting added image (shadow.bdf) is normalized to 1. It optionally can be kept on disk. It corresponds to the fractional "illumination" of the field of view, i.e. a value of 0.9 means that the corresponding point on the sky was obscured by the ribs 10% of the time. See also the 'caveat' below.

For producing the final 0,1 wobbled mask, a simple REPLACE/IMAGE is applied to shadow.bdf. All values above a specified threshold are set to 1, all values below are set to 0. The threshold can be specified on the command line. It is defaulted to 0.95. That means that for the default threshold all areas will be masked which have been obscured by the ribs by more than 5%.

A caveat concerning the source detection

The mask itself works very well when used in source detection. But the intermediate image (shadow.bdf) showing the fractional "illumination" has to be taken with a grain of salt. In principle, such an image could be used to determine the percentage rib obscuration of a source and correct the countrates retrieved from the source detection for that effect.

There are two reasons which object against that:

• The ML source detection (DETECT/MAXLIK, DETECT/SOURCES) takes only the general vignetting pattern into account, but not the vignetting by the ribs. Since the point spread functions are extended and the amount of rib vignetting may vary over the source extent, the resulting count rate or flux and also the source position can be corrupted.
• The used rib maps monew_b.bdf and monew_c.bdf are only approximations. They are 0,1 maps of the ribs, but in reality the rib vignetting does show a more complicated structure.

Summary

If you run ML source detection on sources partially obscured by ribs, you can use the intermediate image shadow.bdf to get a rough estimate how much photons have been obscured by the ribs. But keep in mind that this is only a rough approximation and the results are far less reliable than for not obscured sources.

DOE/MPE 950602

© X-Ray Group at MPE (group)
last update:1998/12/07 12:39:06, editor of this page:Stefan Döbereiner