Flat correction is wrong (polar caps). Feature request: ability to tune or perhaps disable it

melix / astro4j

Astronomy libraries for Java

Apache License 2.0

27 stars 4 forks source link

Flat correction is wrong (polar caps). Feature request: ability to tune or perhaps disable it #364

Closed vnp85 closed 1 month ago

vnp85 commented 1 month ago

Version 2.6.3 dev

Because the polar capping has been observed, I compared the Helium I D3 images generated by JSol'Ex and INTI to exclude an instrument artifact. INTI (on the right) generates no such artifact.

PS: the bad spatial resolution is, in part, a result of the vibrating building. The adjacent building operated a heavy rolling crane.

Line detection seems fine, and there are no overexposed pixels according to Sharpcap Preview. I have a known reflection issue, a reflection that appears and gets worse when the Sun starts to shine through the slit, and the slit's edge is also visible (but avoided by the disk).

vnp85 commented 1 month ago

The issue seems to appear before the tilt is handled:

vnp85 commented 1 month ago

Isn't that flat, by any chance, rotated 90/270 degrees?

vnp85 commented 1 month ago

At the edge detection step, the disk still looks decent, no polar caps. The manually defined line here is the 447.15nm + 45pixel, hunting the He I at 447.15nm

vnp85 commented 1 month ago

As I said above, isn't it possible the flat got applied 90 degrees rotated?

melix commented 1 month ago

No. I don't want to make processing too complicated. If there's something, it should be automatic. The flat correction assumes that the bands are horizontal. Now, you can try to use scripts to achieve what you want, in addition with the fix_banding function which has more options (e.g you can rotate the image, then apply banding correction, then rotate back).

vnp85 commented 1 month ago

Fair enough.