black image after drizzle integration

[javier_laina]

Hi all,

First time I've had issues with drizzle integration process 

After calibrating, cosmetic correcting, wheighting, aligning and integrating 105 Messier 97 H-alpha images, I've tried drizzle integration (as usual in my processing path), and all I get is a totally black image, with "0" value in all its pixels.

Until now, I've used drizzle integration several times without any problem. The last time was processing ngc5371 http://www.javierlaina.es/IMAGENES/ngc5371.html, without any PI updating after it.

I've uploaded 10 M97 xisf images, with its local normalization and drizzle files, so there are 30 files total:
https://www.dropbox.com/sh/qs6unbkronft2or/AAA7Tu3BHqYes09RFBIzJ0ZJa?dl=0

Please, would you kind enough to give a try?

Regards,
Javier
http://www.javierlaina.es/indexeng.html

[John_Gill]

Hi,

Perhaps you have used the wrong data-set.  Did you get a good result from the ImageIntegration and were the Drizzle files updated?

This is not part of the problem, but I often find that I get better results when not using LocalNormalization.

space is not black
John

[Juan Conejero]

Hi Javier,

The data you have uploaded cannot be used for drizzle because it lacks the unregistered images, that is, 'm97_Halfa_000002421_c_cc_w.xisf' (without the '_r' name suffix) and so on.

As for local normalization, it should only be used when it is necessary, not on a regular basis. Unless you have illumination variations that can be fixable with LN (for example, light pollution gradients of varying orientations and intensities), or some other special cases, LN is normally unnecessary.

[javier_laina]

Hi all,

Thank you for your reply.

John, I've got an good image integration previously to try drizzle integration.
I almost always use LN because my observatory is in the middle of the light (what a tittle for a summer song!  ), with best nights' MALE about 4.5, so that I have to fight with all type of LP gradients.

Juan, I've just upload the unregistered images  https://www.dropbox.com/sh/qs6unbkronft2or/AAA7Tu3BHqYes09RFBIzJ0ZJa?dl=0

Regards,
Javier
http://www.javierlaina.es/indexeng.html

[Juan Conejero]

Hi Javier,

Thank you for uploading the images. I can reproduce the same black images with your .xdrz files, with and without local normalization, with and without spline alignment. Obviously, these files are not valid for some reason, which should be related to the way you generated them. I don't know.

I have registered the 10 images and integrated them to regenerate the .xdrz files. DrizzleIntegration works fine with the newly generated drizzle data, yielding a very nice image. I cannot know why this has not worked with your files, but anyway, I would just repeat the whole process carefully because there is no reason for it to fail.

When you apply LocalNormalization, please inspect the normalized files before continuing. Being the weakest part of the process (because of its ill-posed nature), LN is the most likely culprit.

Please let us know how it goes.

EDIT: I have just tried with local normalization data, which I have generated from the registered files, and drizzle integration has worked perfectly. So the problem must be elsewhere.

[javier_laina]

Hi all,

Juan, thank you very much for your dedication.

Well, it seems obvious that I did a wrong step when aligning or integrating the images, so that drizzle files got useless.

I've just done all calibrating and preprocessing steps, and I've have now a well-corrected drizzle image  .

Of course, I'll follow your advice about local normalization.

[javier_laina]

Hi all,

A happy end for ihis issue 

You can see the full processing M97 image:



large image and details: http://www.javierlaina.es/IMAGENES/M97.html

[Juan Conejero]

Javier,

Absolutely fantastic, as well as the rest of images you show on your website. I am very happy to know PixInsight has been helpful to make these wonders!