Pelican & Dark Nebula

[johnrt]

Here is a work in progress project & also my 500th post!!

It is a 4 panel mosaic of the pelican with the rift of dark nebula thrown in for good measure. The final plan will be to include OIII & SII to make a HST colour image.

I have so far completed the Ha frames and stitched them together successfully with PixInsight.

Harry's tutorial video recommends using the DBE tool on individual frames before stitching - but I found this gave me bad seam lines between frames - I'm not saying Harry's video is wrong in any way, but experimenting without DBE'ing the frames worked for me - the frame adaptation process seems clever enough to do the job!

Image details;

Imaging scope & camera - Megrez 72 & Atik 314l+

Guiding scope & camera - ST80 & QHY5

Mount - HEQ5

Filter - Astronomik 12nm Ha

Calibration, integration, stitching & processing all in Pixinsight.

Thank you for looking!

[Telrad]

Absolutely brilliant work in progress of the Pelican and some of the dark nebula, fantastic detail and resolution. Great contrast also. I look forward to the completed work! Thanks for posting

[Catanonia]

Nice work, lots of us doing this mosiac at the moment.

Lovely detail and interesting your thoughts on pre / post DBE as I have been doing mine the traditional way.

[johnrt]

Nice work, lots of us doing this mosiac at the moment.

Lovely detail and interesting your thoughts on pre / post DBE as I have been doing mine the traditional way.

There are indeed some amazing works in progress on this area going on at present.

Has DBE'ing the frames worked for you then? I have to admit I was surprised to find that not using it actually improved things!

Absolutely brilliant work in progress of the Pelican and some of the dark nebula, fantastic detail and resolution. Great contrast also. I look forward to the completed work! Thanks for posting

Thank you very much indeed!

[Catanonia]

There are indeed some amazing works in progress on this area going on at present.

Has DBE'ing the frames worked for you then? I have to admit I was surprised to find that not using it actually improved things!

Thank you very much indeed!

I calibrate, align and then combine each pane and do a DBE on them individually.

I then use star align with mosiac and masks to do pixel math on them to bring them together.

However as you have said, some times getting the last piece in with 2 edges to match can be a pain and I am definately going to have a quick play using your method.

Can you post your PI workflow so I can give it a go please.

[harry page]

Hi

You will probably find with HA images there are very little in the way of gradients if you have applied flats hence your success

If all images are flat without gradients you do not need DBE but most broadband images will have these and removing these perfectly can be difficult

Regards Harry

[johnrt]

I calibrate, align and then combine each pane and do a DBE on them individually.

I then use star align with mosiac and masks to do pixel math on them to bring them together.

However as you have said, some times getting the last piece in with 2 edges to match can be a pain and I am definately going to have a quick play using your method.

Can you post your PI workflow so I can give it a go please.

No problem..........

I calibrate the images using flats and bias. Then register using star align, integrate (median combine with pixel rejection) and crop out any dodgy edges.

I then used the star alignment tool with register/union mosaic and the frame adaptation box checked on the un-stretched images - building up the image frame by frame.

I found it best to start off with the poorest frame first (if you have one) and stitch on to that first. It is sometimes a little bit of trial and error on which image to use as reference and which to apply to.

When the image was combined I re-cropped and applied the DBE then processed as normal (stretch, wavelets, noise reduction etc).

[johnrt]

Hi

You will probably find with HA images there are very little in the way of gradients if you have applied flats hence your success

If all images are flat without gradients you do not need DBE but most broadband images will have these and removing these perfectly can be difficult

Regards Harry

Harry,

Does the frame adaptation work off matching the background?

[harry page]

Hi

It works by matching the background and signal areas i.e 2 no adaptions.

It does however work better when there is a reasonable amount of background for PI to evaluate the corrections

There is another mosaic tool coming which will be even better , but will have to wait a week or 2 for this

Harry

[Catanonia]

Hi

It works by matching the background and signal areas i.e 2 no adaptions.

It does however work better when there is a reasonable amount of background for PI to evaluate the corrections

There is another mosaic tool coming which will be even better , but will have to wait a week or 2 for this

Harry

Tease, eagerly awaiting this one from PI

Useful thread and will definately have to play a bit with my mosiacs. Lucky I am still capturing frames.

Just set up for another panel or 2 across the top of mine now

Waiting for darkness... grrrr summer lol, at least in shorts

[Catanonia]

No problem..........

I calibrate the images using flats and bias. Then register using star align, integrate (median combine with pixel rejection) and crop out any dodgy edges.

I then used the star alignment tool with register/union mosaic and the frame adaptation box checked on the un-stretched images - building up the image frame by frame.

I found it best to start off with the poorest frame first (if you have one) and stitch on to that first. It is sometimes a little bit of trial and error on which image to use as reference and which to apply to.

When the image was combined I re-cropped and applied the DBE then processed as normal (stretch, wavelets, noise reduction etc).

mmmm, not sure if I have been using frame adaption. I have been using the video, might have been Harrys on using Pixel Math to balance each frame.

Thanks for the workflow, so much PI can do, only just really getting into it now after about 4 months of heavy use.

[johnrt]

but will have to wait a week or 2 for this

Harry

We have to WAIT!? Aaaarghh!

[Catanonia]

HAd a quick play with Frame Adaption and seems nice. Takes some of the guessing work out of Pixel maths.

I got 2 more panes last night, so will have a play with DBE after stitching

[MartinB]

You've captured the dark nebulosity beautifully John.

[johnrt]

You've captured the dark nebulosity beautifully John.

Thank you - I think the dark nebula area that splits the NA & Pelican is much more interesting than either of the 2 more recognisable structures. A bit odd maybe?

The frame adaptation tool is a clever little bit of programming.

[Catanonia]

Tried your method of non dbe before stitching and didn't really work for me.

BUT

Thanks for the frame adaption check box, that is really really useful mate.

[johnrt]

Tried your method of non dbe before stitching and didn't really work for me.

BUT

Thanks for the frame adaption check box, that is really really useful mate.

You're very welcome!

Weird that the DBE makes things worse for me, not sure I understand why. I'll see what happens when I collect the OIII & SII - should be easier as there will be more background for the frame adaptation to work off.

Just out of interest is it best to colour the frames individually then stitch or stitch 3 channels separately then combine to make the colour image?

[minte222]

Looks stunning cant wait to see this?

[johnrt]

I've finished added the OIII to my 4 panel mosaic - stitching those frames together is a challenge.

Sky conditions vary wildly from night to night and the OIII suffers from this so much more than Ha.

I had to re-shoot the bottom section 3 times and the bottom right panel is still too noisy for my liking but I'm not doing it again! It stays as is.

Up to 8 hours exposure - all the leg work done in Pixinsight.

Thanks for looking

original size in my flickr stream in signature.

[ollypenrice]

I really admire the Ha version but find the HaO111 a bit magenta...

A small thing but you do have some dark halo effects around stars. Could this come from the sharpening? Maybe you could have a star layer to protect them?

Anyway I agree with you about the dark nebulosity being so interesting. Have you come across the tool in PI utilities called Enhance Dark Structures? It's excellent but you do need a lot of signal for it to work. You have some great contrasts in this picture.

Olly

[johnh]

This is impressive work and the dark areas certainly come to life.

John.

[johnrt]

Thanks Olly & John.

The dark halos could be from either the sharpening or the star reduction in PI, they were masked for both processes but I might have been a little aggressive there - I like little stars!

I also use the Dark structure script as a matter of course - it's a brilliant bit of kit!

[johnrt]

I've replaced the colour version with a better red balanced image tweaked in CS5. Last version very purple!

[Coco]

I very much prefer the first one, I know its a 285 chip set but have you used a smidgen of noise reduction like ACDNR? it just looks too smooth on my monitor..

That said I really like the marble effect synonymous of this area that this image has captured.. On my big dell monitor it looks great , on my HP lappy with glossy screen it looks over noise reduced

The Ha mono gets my vote and its a cracker

[johnrt]

I very much prefer the first one, I know its a 285 chip set but have you used a smidgen of noise reduction like ACDNR? it just looks too smooth on my monitor..

That said I really like the marble effect synonymous of this area that this image has captured.. On my big dell monitor it looks great , on my HP lappy with glossy screen it looks over noise reduced

The Ha mono gets my vote and its a cracker

Yes I know what you mean - there seems to be a lot of detail lost from the Ha version - and your comment rung a bell. I had been messing around with wavelet noise reduction the mono Ha image before combining - and had forgot to undo it.

So I went back and re-processed a third time!