Q: What do the values in the mask file delivered with my data represent?
A:
There is a description of this data at the bottom of the accompanying
.hdr file for the mask files. The values are:
| Value | Description |
| 0 | Good data |
| 1 | Underflows |
| 2 | Overflows |
| 4 | Bad pixels in the CCD detector |
| 8 | Pixel affected by uncorrected smear |
| 16 | Dropped scans |
| 32 | Corrupt raw data |
| 64 | Quality control failure |
If there is more than one issue with a pixel it will have a value equal to the sum of the corresponding values. For example, a bad pixel which has also overflowed would have DN=6 (2+4).
The bad pixels in the CCD detector (i.e. the pixels marked 4 in the mask file and in the separate ASCII file) come from a list supplied by the manufacturers resulting from the Hawk instrument calibration. These pixels are supposedly not recording accurately and are not to be trusted. We know this list to contain certain inaccuracies and hope to generate a more representative list in the future.
EDIT:2012 Data that is undergoing re-processing (and new data from 2012 onwards) will use a newly generated list created from the ARSF calibration data.
The masks have been generated as a by product of applying the calibration to the raw data. So if during calibration the value of a pixel has underflowed (gone below 0) then we give it a value of 1 in the mask file. Similarly, if the data has overflowed (> sensor maximum) we will apply a value of 2 to the mask.
In the level 1 data, underflowing pixels are set to 0 (the minimum possible for an unsigned integer). An overflowed level 0 pixel may result in a valid but wrong level 1 value, which is kept if it is less than 65535 (the maximum allowed level 1 data value) or otherwise forced to 65535. We retain the computed value mostly in order to allow people to quickview the image more easily without scaling problems, but it should be masked when actually used.