Image Integration of data from 2 separate nights? Can't get the images to align properly?

[oymd]

Hi everyone

I've been imaging IC1318 over the past 2 nights.

After my first night of imaging, I removed the camera and cleaned the optics.

So, for the second night I took a new set of Flats and dark flats.

I've used NINA throughout, with a Manual rotator, so both sets of LIGHTS are pretty much the same with exactly the same orientation.

I did the WBPP for EACH set of Lights with its corresponding calibration frames.

In WBPP I did NOT click image integration.

As such, I'm left with TWO sets of REGISTERED files, each from a separate night, and each calibrated with its own flats, DFs. Darks are COMMON to both.

For some reason, in IMAGE INTEGRATION, when I add ALL the registered files together, the final image is not right? The two sets are MISALIGNED?

What am I missing?

Image Integration on EACH set of Registered lights works fine if done separately, but once I COMBINE ALL the lights into Image Integration, its all misaligned?

Everything is on DEFAULT settings. I tried WITH and WITHOUT pixel rejection?

Here is Image integration of the first night set of registered files:

 

Here is Image integration of the second night set of registered files:

 

And this happens if I add BOTH sets to Image integration?

Thanks

[scitmon]

You need to run a Star Alignment on your final images from each night.  Once you've done this you can combine the two images with pixel math or image integration.  If I remember correctly, image integration needs at least 3 images to integrate, but you can get around this by cloning each nights image, then combine the 4.  

[oymd]

1 hour ago, scitmon said:

You need to run a Star Alignment on your final images from each night.  Once you've done this you can combine the two images with pixel math or image integration.  If I remember correctly, image integration needs at least 3 images to integrate, but you can get around this by cloning each nights image, then combine the 4.  

Thanks for your reply.

I ran the images through star alignment, then ran them all in Image integration. 

I do not understand what you meant by running at least 3 images?

My data consists of 2 sets, each 47 images.

After running WBPP on each set separately, I have 2 sets of REGISTERED images.

they look like this: 2021-10-07_19-49-19__-4.90_300.00s_0005_c_cc_d_r

I then ran them ALL together in Star Alignment, and for the reference frame, I picked the frame that was already chosen by WBPP as the REFERENCE IMAGE.

That added an ADDITIONAL _r at the end, the the files looked like this: 2021-10-07_19-49-19__-4.90_300.00s_0005_c_cc_d_r_r

I then added ALL the files into IMAGE INTEGRATION, and now I have a working image.

But I think I'm not doing this right.

Since I have TWO _r symbols at the end of my images' file names, I think I'm doing registration TWICE? Or is it normal for Registration to add a _r and then Star Alignment to add another _r?

Maybe in WBPP I should NOT click Registration at all, so I am left with CALIBRATED FILES only, and the do Star Alignment on those, and then Image integration finally?

[oymd]

Ok, So I just repeated the whole process, and I think this is the correct way to do it?

Not sure?

Ran WBPP on EACH set of 47 images separately.

In the lights tab, I DID NOT CLICK THIS TIME ON EITHER IMAGE REGISTRATION or IMAGE INTEGRATION.

I think Image Registration option in WBPP is actually Star Alignment??

So, I'm left with TWO sets of debayered, cosmetically corrected sets of images, ending in c_cc_d

I ran those 2 sets in Star Alignment, and then finally the resulting images in Image integration?

Is that right?

[scitmon]

I'm not overly familiar with WBPP as I learned to do it the manual way, but yes I think you are right that image registration does indeed try to star align.

I've done a little googling and turned up something interesting... for each batch, are you picking separate reference frames?  If so, try picking the same reference frame for both batches.

Edited October 11, 2021 by scitmon

[Richard_]

In your case you ran the WBPP twice using different sets of data. The script doesn't know to align the subs from the second night to the subs from the first night. We can tell the WBPP to do this but it means you need to put all files from both nights into the script at once, which calls for good filenaming. 

Using the WBPP script you can stack all your subs across multiple nights by using a keyword(s). In your case, you want to calibrate your subs by date (because you will have different flat frames each night) but you will want to register subs from both nights together and create a single master light. Register is just a fancy name for aligning your subs. Integrating means to combine the subs together. 

Adam Block explains how to do this extremely well in one of his videos, I'd highly recommend you check it out (link below). Using this approach, I was able to calibrate subs from multiple nights and output a separate master light for each panel of a mosaic I was working on using a single instance of the WBPP script. This was all made possible by using keywords in the filenames and utilising them in the WBPP script. 

 

Edited October 19, 2021 by Richard_
Correct some phone typos :)