Workaround for AIA dark images

31 views
Skip to first unread message

Felipe Meza

unread,
Feb 21, 2024, 6:03:20 PMFeb 21
to su...@googlegroups.com
Hi, recently some images acquired are darker than usual, for instance:
img_sol020.jpeg

Any idea about how to fix this?

Regards,

  
Message has been deleted

Albert Y. Shih

unread,
Feb 21, 2024, 11:24:43 PMFeb 21
to su...@googlegroups.com
Hi,
     The dark frames you are seeing are short AIA exposures during an X-class flare when using its automatic exposure control.  In this frame, the exposure is only ~0.18 seconds, which is far shorter than the normal exposure time.  The default normalization of sunpy plots will have the pixel brightnesses span from the lowest data value to the highest data value, so under quiet conditions, a short exposure would look quite similar to a long exposure.  However, during a bright flare, data values will saturate on exposures of all lengths.  Given the default normalization, this makes the non-flaring pixels look much darker relative to flaring pixels on short exposures.
     To get the plot to look like a typical frame, you need to clip out the flaring pixels.  Here are two ways to do so:

1) You can use the `clip_interval` keyword, as shown in this example (https://docs.sunpy.org/en/stable/generated/gallery/plotting/aia_example.html), to have the normalization ignore the most extreme percentiles of data values.  Here's the same frame with the top 0.05% of values clipped out (and the bottom 1% clipped out) by calling:

`.plot(clip_interval=(1, 99.95)*u.percent)`
Figure_1.png

2) You can manually set the `vmin` and `vmax` for the normalizer for the plot:

`.plot(vmin=0, vmax=400)`

The units for `vmax` is in DN, i.e., the digitized value of the total flux in each pixel, and thus depends on exposure time.  If you want to try to match the looks of plots of a short exposure and a long exposure, the ratio of their `vmax` values will be (approximately) the ratio of their exposure times.

Hope that helps!

Albert

--
You received this message because you are subscribed to the Google Groups "SunPy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sunpy+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sunpy/CADhWRJAoR_901occRoits%3DugBKg0oor0hW2GGAGNaWkKVu4dcQ%40mail.gmail.com.
Reply all
Reply to author
Forward
0 new messages