Image statistics process

[astrodoc71]

I'm using the image statistics process to measure percent saturation of a spectrum by looking at maximum ADU for the image (see attached). This is a raw image, no dark subtraction. Since there is no calibration at all, is this going to be an inaccurate measure if there are hot pixels present? Or is this process not really affected by hot pixels?
Thanks
Dave

[bulrichl]

The maximum of the whole image is certainly not a good criterion; at worst it might represent just one hot pixel. I would crop the background area and take a look at the histogram (in conjunction with the statistics) of the cropped image: the decreasing shoulder of the histogram peak should be a much better estimate.

Bernd