Nebulosity Fits Geometry issue

[cdavid]

Hi, I captured some images the other night with Nebulosity.  I brought them into pixinsight for batch preprocessing.  When it attempts to calibrate the Light frames with a superbias I had previously made in Pixinsight I get an error stating "incompatible image geometry"

I think it has to do with the way Nebulosity saves the FITS header but I am not sure what to do.  Anyone have any ideas?
Thanks
Carlos

[cdavid]

...I looked at the fits header for Nebulosity images compared to my pixinsight Superbias image.

The Pixinsight image is 3906 x2602 and the Nebulosity Fits is 3906 x 2604

I'm clueless as to why this would be???

[cdavid]

anyone?

[Geoff]

Accidently had a subframe selected when taking the bias frames? Check the size of your original bias frames.
Geoff

[cdavid]

Thanks....I went back and checked all my previous bias darks and some lights. The PIXinsight fits files are the same as my raw 3906x2602. The Nebulosity acquired Fits are for some reason 3906x2604.

There is something strange going on when Nebulosity saves its files. I can't figure this out. Essentially I cannot calibrate this set of subs as they were acquired and saved by Nebulosity and the flats that night were directly obtained by the camera.

[msmythers]

Just a thought but what about resampling the flats by the 2 pixels and see what the results are like?



Mike

[cdavid]

nope...it changes the other dimension too when i tried that

thanks for suggestion!

[msmythers]

Did you have the 'Preserve Aspect Ratio' unchecked?

Here is an example. I changed the height only by 2 pixels.



Mike

[cdavid]

Duh! I'm a dumb ass. I'll try that. Thanks

[gordonewen]

David, I have been having the same problem as you, but today all worked for some unknown reason.
I have a qHY10 I was using on NGC7331 acquiring images with Nebulosity 3. I had a set of 20 light frame's, 50 dark frame's, 25 flats, 25 flat darks and a set of bias frames. All the lights, darks, flats and flat darks had pixel dimensions of 3896x2612, but my bias frames had 3894x2610. Very odd, all taken with same camera and pc. Today I tried a 0.001s exposure in lieu of a bias frame of 0.00s exposure. It gave the correct pixel dimensions. I then tried 0.0001s, it also gave the correct dimensions. For the hell of it, I repeated the 0.000s exposure..................... I got the correct dimensions. Why I got the correct dimensions today mystifies me, but all is well. As I have said elsewhere, as son as we get a PI image acquisition module the better!!!
Have you repeated the exercise to see if you get the same result?
Regards
Gordon

[cdavid]

Hi Gordon

This is Carlos....from the other forum.....we were discussing this over the weekend.  For the benefit of anyone reading this I will summarize the issue.

I am using a DSLR and when I use Nebulosity 3 to capture the images it gives the option to save to the computer as a FITS, FITS and CR2 to the Computer, FITs to Computer and CR2 to CF card in the camera, and CF card only.

I have my camera a Canon 1000D set to save to the Computer as well as to the CF card.  During one of my sessions recently I saved as above, but I prefer to shoot my Flats using the camera shutter button and the camera set to AV.  I have not been happy trying to determine the best exposure for flats using the settings in Nebulosity hence why I shoot this way.  Further I have a library of BIAS and Darks that have previously been obtained with the camera shutter button as well.

So when I went to process the images with my library BIAS, Darks, and the new Flats (all which have been taken via the camera shutter and saved as raw CR2 files to the camera CF card they did not match in size the FITS saved to the computer.  Normally I would not have noticed because I usually just use the images saved to the Camera CF card and bring them into Pixinsight.

This time however, I had glitch while setting up Nebulosity and it did not save my light subs to the camera and only to the computer as FITS. 

I could not understand why there was a difference in image size.  I emailed Craig Stark but have not been able to determine why or if there is a setting to fix this.  In any event, analyzing the FITS images it became apparent that there were two extra rows of black pixels on the bottom of the image.  Hence my difference in geometry (3906 x 2902 for the normal CR2 images on the card and 3906 x 2904 for the FITS saved to the computer).

I have been able to salvage this set of images by creating an image container, using the crop tool to remove the bottom two rows and applying to all the images in the container.

I still do not know why this is happening, but this is the fix if anyone out there runs into the same issue.
cheers
Carlos