next up previous contents
Next: Construction of Flat Fields Up: Off-line Reduction Previous: Off-line Reduction

   
Bad Pixel Detection and Removal

Before most of the reduction process can be conducted, bad pixel values must be removed. This is usually achieved by flagging pixels above or below some threshold value as bad, and replacing their values by those of nearby, non-bad pixels. The LAMP ON or LAMP OFF flat field images are a good place to start to define your bad-pixel map. To determine the threshold values to use, you should examine the statistics of the image and set thresholds 5 to 10 standard deviations away from the mean in both positive and negative directions. The exact thresholds for determination of bad pixels will depend on the details of your flat field observations - filter, objective lens, integration time and lamp voltage - so you should experiment with different values until you obtain a satisfactory result.

Because IR arrays are sensitive to thermal cycling and to atmospheric contamination, the bad pixel lists change over time. Recent lists are available via the ESO WWW pages for comparison to bad pixel maps you generate yourself, but are unlikely to perfectly match the list derived from your own data.

In MIDAS, bad pixel detection is possible with the command MASK/IRAC2; the command CMASK/IRAC2 can be used for bad pixel removal. Refer to the MIDAS manual for further documentation of these commands.


next up previous contents
Next: Construction of Flat Fields Up: Off-line Reduction Previous: Off-line Reduction
Petra Nass
1999-06-15