3.2.6 Subtleties of Background Subtraction


next up previous contents
Next: 3.2.7 Charge Bleed-back Up: 3.2 The CCD Camera Previous: 3.2.5 SAA Effects on

3.2.6 Subtleties of Background Subtraction

In general, the background subtraction provided by SXT_PREP, DARK_SUB, and LEAK_SUB work amazingly well for such a complex problem. If background subtraction is a little off, substantial errors can accumulate in computing the intensity of large areas of faint emission. To better match exact exposure times the keyword /dc_interpolate should be used. For cases where even this is not satisfactory, the keyword /shift_floor will use the average of a area in the lower left corner of the image as the background. This is usually only needed for the very shortest exposures.

Note that these routines leave the negative and zero values remaining from dark frame subtraction. These may need to be replaced with a positive value for, e.g., logarithmic display or ratio forming. By default SXT_TEEM will ignore negative intensity pixels when computing temperatures.

Dark subtraction will not remove cosmic or SAA artifacts which can be treated manually by replacing the offending pixel with the average of the surrounding pixels or through the use of DE_SPIKER.   

After Nov-92, some background due to a visible light leak also needed to be removed. See section 3.3.5 for more details. 


ydac@mssly1.mssl.ucl.ac.uk