Show correlation coefficient or covariance of noise in overscan region

[julienguy]

Once we have the correlation coefficient of the noise across ~10 pixels saved in the processed image headers
(see desihub/desispec#2173), it would be useful to display the results in nightwatch.

For instance, it's hard to tell whether the correlated noise in the following image is higher than usual or not:
https://nightwatch.desi.lbl.gov/20240130/00222364/preproc-z9-00222364-4x.html

[sybenzvi]

Assuming the correlation coefficients are available at the per-amp level, we can add this information to the CCD QA plots after establishing typical ranges.

Is this something we want to show for any exposure, or just certain ones (like darks)?

[julienguy]

We could show it for each exposure