Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://xmm.vilspa.esa.es/sas/8.0.0/doc/ommodmap/node6.html
Дата изменения: Wed Jul 2 09:51:59 2008 Дата индексирования: Fri Sep 5 19:42:13 2008 Кодировка: Поисковые слова: п п п п п п п п п п п |
The problem relates to the distribution of events, not to changes in the sensitivity, and so dividing the science image by the mod-8 map, whilst providing a simple cosmetic solution, would compromise the photometry of the image. Instead the algorithm resamples the image to give each image pixel equal area within a mod-8 tile.
For each mod-8 tile in the mod-8 map, the pixel x-boundaries are determined such that within each row of pixels the counts/unit pixel area is the same for each pixel, as illustrated in the left hand panel of Figure 1. This is then taken as the spatial layout of the pixel x-boundaries in the same mod-8 tile in the science image and the science image is resampled to a grid of evenly spaced pixels. Then the y-boundaries of each row are determined such that each row has the same number of counts/unit pixel area, as illustrated in the right hand panel of Figure 1. This is then taken as the spatial layout of pixel y-boundaries in the same mod-8 tile in the science image, which is then resampled to a grid of evenly spaced rows.
The same procedure is then repeated with the order reversed (remapping in y followed by remapping in x) and the final corrected science image is made from an average of the two resamplings. This is to ensure that the redistribution of photons is performed in an identical fashion in x and y.
The threshold for applying the mod-8 mapping algorithm is currently
established to be 1000 mean background counts from the nBox * nBox area,
i.e. the mean background value times nBox must exceed 1000 counts.
The motivation for doing that was due to the fact that
the mod-8 noise up to 10 to 20% assumes that in order
to improve the image we have to have an accuracy
of about 3%, which corresponds to mean*nBox
> 1000 counts.
The algorithm is not applied to images with poor statistics.