European Southern Observatory New Technology Telescope

General :
Observations :
Operations :

The realization of a bad pixels mask in SOFI is done anually since september 1999. There are two types of observations to do : flats and darks.

FLATS :

Check the setup of SOFI :

• Large_Field
• filter Ks
• open
• open
• LF

Typically, the DIT is 2.0 and the NDIT is 5. Choose a name for the flat exposure in the expo_name field of the OS panel. Verify that the count number is around 4000 ADU.Take 30 flats.

DARKS :

Put the instrument in DARK mode and change the name of the exposure.Take the same number of darks than flats.

The data are stored @wsofi $INS_ROOT/SYSTEM/DETDATA.They can be reduced using bad_pix.cl  script in IRAF.

The different steps are :

• imcombine the flats to obtain the average (--> flat_average) and the sigma (--> flat_sigma)
• imcombine the darks to obtain the average (--> dark_average)
• Plot the histogramme of each of them (imhist, see an example). Choose windows that are considered as "good pixels" and "bad pixels". To do that the procedure fits a gaussian in the distribution and determines the mean and the fwhm (see the splot result). Reenter the values in the task.
• Replace (imreplace) the good pixel by 0, and the bad ones by 1. So the distribution has only two pics.
• There are three different types of bad pixels : dead, noisy,hot. Average them (imcombine with no weigth --> mean). In this total average, the values are 0,1/3,2/3,1 (see the histogramme). All the pixels with an intensity above or equal to 2/3 are bad (so imreplace them to 1) and the others with an intensity under 2/3 are good (so implace them by 0). It's the final pixel mask (see the histogramme of the mask).
• Determine the number of bad pixels, just doing an imstat with the lower value equal at 0.8 and the upper value at 2. Normally, the percentage of bad pixels should be lower than 1%.

   

Send comments to: NTT Webmasters <nttweb@eso.org> Last Update: January 17, 2004 (Malvina Billeres)