Image Analysis

At this point the user is left nearly to their own devices. For scientific analysis the data should not be smoothed and, depending upon the analysis needs, binned. A very simple task, binadapt, is provided to combine all the components properly to create and image, perhaps with binning. (In the next section we'll deal with the same routine in terms of smoothing.) The output file is in count rate:

$\displaystyle R=\left(\frac{\sum_{inst}c-[\sum_{inst}b+\sum_{inst}p+\sum_{inst}s]}{\sum_{inst}f(\alpha)e}\right)$

(1)

and count rate uncertainties:

$\displaystyle \sigma_R=\left(\frac{\sqrt{\sum_{inst}c}}{\sum_{inst}f(\alpha)e}\right)$

(2)

where $f(\alpha)$ is a scale factor (as a function of the power law index) used convert the exposure maps from a particular instrument, filter, bandpass combination to the same bandpass for the MOS2, Medium filter combination.

to units of counts s $^{-1}$ deg $^{-2}$ , and is equal to the reciprocal of the fundamental analysis pixel size of $2.5''\times2.5''$ ( $=2.0736\times10^6$ ). $c$ is the number of counts, $b$ is the number of model particle background counts, $p$ is the number of model SPF counts, and $s$ is the number of model SWCX counts). The assumption is that the counting statistics of the observation dominates the uncertainty in the model background counts. In practice this probably underestimates the uncertainty which would make any analysis results more conservative. The combination of total counts, backgrounds, exposure, and mask is done with binadabpt.

binadapt prefix=comb elow=400 ehigh=1250
withpartbkg=true withspbkg=true
withswxcbkg=true withbinning=false
withsmoothing=false

As might be divined from the above, binadapt can be used for a single detector (for which the prefix would be something like mos1S001), or for the combination of any of the three EPIC detectors (in which case the prefix is comb). The elow and ehigh are used to construct the input file names. The withpartbkg, withspbkg, and withswcxbkg flags can be set to include or exclude the various backgrounds as required. Even if a mask is not specified, one can set a minimum exposure threshold, maskthresh to exclude strongly vignetted regions. Of course, an extra mask can be used as well withmask=true maskfile= and the name of the mask file.