XMM-Newton Science Analysis System


esplinemap (esplinemap-5.4.1) [xmmsas_20230412_1735-21.0.0]

Description

Sources found in the local detection step (task eboxdetect) at significance levels (column SIGMA of eboxdetect source list) exceeding a user-specifiable threshold (input parameter mlmin) in the respective energy band are removed from the image using a suitable PSF and source brightness dependent cut-out radius (determined to be the radius at which each source contributes more than a user-specifiable number of counts/arcsec$^2$ to the background; parameter scut; default value: 0.01). The resulting image can optionally be output for diagnostic purposes. After the removal of the sources the image is rebinned to a grid of the dimension nsplinenodes $\times$ nsplinenodes.

Division of the image by the exposure image removes gradients due to spatial variations of the exposure from the image which otherwise would not be well represented by the spline fit. A two dimensional spline fit of the rebinned and exposure corrected image is performed. The number of spline nodes (default value: 16) is user-selectable. Finally, the resulting spline image is again multiplied by the exposure images. If the parameter nfitrun is set to values $>1$, remaining excesses of the input image over the result of spline fit can be removed iteratively: if pixels of the rebinned image deviate from the spline fit by more than a specifiable number of sigmas (default value: 4 sigmas) the excesses are removed by setting their statistical weights to zero and the spline fit is repeated (maximum number of iterations may be specified). The number of removed bins and the reduced chi$^2$ values are displayed when using verbosity level 5 or higher. Note that removal of a large number of contiguous bins will lead to areas where the spline fit is unconstrained.

The reduced chi$^2$ and corresponding number of degrees of freedom of the background map with respect to the input image is stored in the keywords CHISQR and NDOF of the output background map.

From version 3.0 esplinemap is able to determine the background caused by out-of-time events registered during the readout process of the PN CCDs. If the flag withootset is set, the photon event table specified in ooteventset is read and the background caused by OOT events is included in the output background map. As input table esplinemap can use either a normal photon event data-set or a photon events table created with epevents with flag withoutoftime set. Note that in both cases a photon event set has to be filtered with the same temporal and flag selections as the image used as input to esplinemap. The parameters pimin and pimax are used to specify the energy range of the input image and to select those photons from the input event list that fall into this energy range. If the input event table contains only photons within the energy range of the input image, the parameters can be left at their default values pimin=1 and pimax=30000.

With version 4.0 an alternative method to fit the background of an image has been implemented: If the option fitmethod=model is set, a 2-component model for the detector (particle) and the cosmic X-ray backgrounds is fit to the masked and binned input image. This model consist of a linear combination of the vignetted exposure map and the unvignetted exposure mask of the input image. The exposure maps are specified by the user via the parameters expimageset and expimageset2. An example call of eexpmap and esplinemap is given here:

   eexpmap imageset=image.fits eventset=events.fits attitudeset=attitude.fits \
           withvignetting=yes expimageset=expmap1.fits pimin=500 pimax=2000

   eexpmap imageset=image.fits eventset=events.fits attitudeset=attitude.fits \
           withvignetting=no expimageset=expmap2.fits pimin=500 pimax=2000

   esplinemap imageset=image.fits boxlistset=eboxlist.fits withexpimage=yes \
              bkgimageset=bkg_model.fits \
              withexpimage2=yes expimageset=expmap1.fits expimageset2=expmap2.fits \
              pimin=500 pimax=2000  \
              fitmethod=model

If only one exposure map is provided, it is assumed to be a vignetted exposure map and a flat image is used as the second model component. All other parameters of esplinemap can be used as in the case of fitmethod=spline.

A third fitmethod has been introduced with version 5.0. fitmethod=smooth invokes an adaptive smoothing of the cheesed background image. It is convolved with an Gaussian kernel whose width is increased by factors of $\sqrt{2}$ in eight consecutive steps. For each image position, the layer with the best signal-to-noise ratio is chosen out of the eight smoothed image layers. Neighboured layers are interpolated to achieve the final smoothed background map. The adaptive smoothing is particularly useful if the background is varying strongly over the field of view and set as default in edetect_stack.