APT Virtual Mosaic,

[Paul M]

Something lit a fuse in my tiny mind recently regarding mosaics! My ASI071 camera isn't the tool for imaging M31 or even M45 as a single frame. I hadn't really thought about it until reading up on the symbiosis available between APT and HNSky. I use HNSky as my preferred planetarium on my imaging computer.

I found a video that introduced a method of creating a frame plan:

That uses CDC but the same feature is available with HNSky. I created a frame that corresponds to my sensor. By positioning the frame appropriately and hitting the home key the frame will persist and allow adjacent frames to be positioned.

Here are 6 overlapping frames on M45:

Right clicking the mouse brings up the option to export the frames via the "server", the connection set up between HNSky and APT. 

Then in APT open "Objects" in the Gear tab and select "custom" and import list" :

Select as shown and the 6 frames from HNSky appear as custom objects that can be used in a imaging plan.

But how to Automate?

My solution uses a recent addition to APT's imaging planner; the ability to add a a script or command. I couldn't see how to execute a command that didn't expect an image exposure being included. So I have 12 entries in my plan. 1 entry to execute a GOTO++ to the coordinated each of the imported frames and 1 entry to execute the exposures for that frame. Here I've had to put a 0 seconds exposure with each GOTO command because it requires something in there to save out that line:

You can see the complete mosaic plan on the right there. My first attempt at this was tested under a real sky and it failed. That's how I found it required a separate line for the GOTO++ command!!

So how to test without wasting a clear night?

From the same stable as HNSky and ASTAP, Han has created an Ascom sky simulator: https://www.hnsky.org/

I'm not going to go into the details of setting that up. It took me 2 full evenings of trial and error and I eventually resorted to reading the instructions, yeah, I know...

Once set up with APT by selecting the correct ASCOM simulators it will simulate a camera pointing at a real or simulated sky. I chose DSS imagery because I wanted to plate solve to prove my whole mosaic plan. It also simulates a focuser, I used that routine just for kicks!

I'd be lying if I said getting this far was a bed of roses. I nearly gave up but I eventually got to a point where I could run my plan and have 6 images to stitch as a mosaic.

Here is an example image from the 6 frames:

To stick with the HNSky family I used one of my favorite pieces of astro software, ASTAP. It has an astrometric mosaic stitching function:

And, AND! when it finished I got this:

I know, it isn't really my image....or is it? 🙈

I've done no other processing outside of ASTAP's automatic background equalisation. There was a big difference in image quality of some of the frames so it worked out better than I expected.

But don't get me wrong, as excited as I am about getting the above result, the real reason for this post is to demonstrate a method of imaging large scale objects using HNSky as a mosaic builder and Sky Simulator to test it.

Maybe it'll be useful to a few others. I've always enjoyed laptop astronomy. Playing with software is a great pastime on cloudy nights.

[Atreta]

Thanks a lot for sharing this, it will be very helpful. 

Great image of the Pleiades too, congrats 

 

[SpookyKatt]

Thanks for posting, I wasn't aware of that capability but its useful to know.

Kathleen

[Paul M]

9 hours ago, Atreta said:

Thanks a lot for sharing this, it will be very helpful. 

Great image of the Pleiades too, congrats 

 

Just to be clear. That isn't really my imaging that produced the final mosaic.

It's just stitched together frames of DSS imagery to demonstrate the imaging plan.

The ASCOM Sky Simulator just made APT think it was imaging a real sky by providing camera frame sized bits of DSS imagery based on where the telescope simulator was "pointing".

But hopefully when I try it on a real sky I'll get something even better!

Edited September 27, 2021 by Paul M

[mackiedlm]

Great write up, thanks. Ime really interested in this but currently I use stellarium and I don't know if there is a way to place more than one frame at a time and then export them all. So I may have to install hnsky just for this.

 

Thanks for the info.

[Paul M]

2 minutes ago, mackiedlm said:

Great write up, thanks. Ime really interested in this but currently I use stellarium and I don't know if there is a way to place more than one frame at a time and then export them all. So I may have to install hnsky just for this.

 

Thanks for the info.

I think that currently APT supports only the import of frames from CDC and HNSky.

I played with CDC even before I started down the Yellow Brick road of imaging and found it "difficult", not impossible just counter intuitive. Then I found HNSky packaged with Astroberry and didn't rate it. I had a play with it on my W10 lappy  and with some effort tuned it to my requirements. Now it's used frequently on my imaging PC and for planning on my lappy.

Maybe APT will eventually include a fully featured mosaic framing module but will likely still call on an external planetarium such as CDC and HNSky.

[Paul M]

All of the above was performed on virtual data, just to prove the method.

Well, Saturday night was clear so I decided to have a go at M45 for real!

It turned out to be a mixed session. While I was focusing I noticed the sky was getting milky, the waxing gibbous Moon was still still affecting things low in the sw and I could see a layer of mist floating past it. I nearly gave up there and then but by the time I was set up the sky was better.

Off I set, I went for 6 frames of 10 x 120 sec, binned 2x2, ASI071MC Pro, Optolong L Pro filter, 250 PDS Newt, guided by PHD2. I took flats the following morning.

The imaging plan worked flawlessly  And, AND! I had my first ever automated meridian flip mid sequence, all done by APT

Now the bad news, the sky was still very variable and as M45 climbed into darker sky the Moon set and Blackpool's Illuminations were put to bed. The prescribed method, which I used, is Stack the subs for each frame separately using normal star alignment and apply flats. Then stack the resultant frames using the mosaic mode. I also used the auto crop function which I set to crop each frame 5% all round to remove artifacts. There is also a background equalising function which I used.

And ASTAP set about doing its thing. It took a while on my old coal fired lappy. 

Yes, I know what you are both thinking, YUK! To me it is a thing of beauty!

I've tried all ways up to fix the joints but there are just too many gradients and artifacts. As usual my flats didn't work perfectly but each individual frame looks fine when processed in isolation. Lots of nebulosity but as always more data would have been better. 

Using wavelets in Siril does remove some of the issues but introduces others that make the image worse.

Anyway,  the method worked brilliantly. I just need a dark sky...

[han59]

Have you tried in ASTAP the option crop images and maybe "Merge overlapping background=off"?

Probably it also works better if you equalize the background of each image before making the mosaic.

For testing/experimenting, if you could share the 4 or 6 raw images, I could try to get it better.

An alternative for making a Mosaic is to use the free Microsoft program ICE which is discontinued but still can be found at some unofficial places.

Han

author ASTAP

 

[Paul M]

Hi Han, first can I thank you for the continued development of the HNSky family!!

I've tried ICE, it just didn't want to know! I did stack the individual frames a few times with slightly different approaches including with and without Merging overlapping background. Each time I had the crop option set.

Here are the 6 frames that I put into the final mosaic above:

M45, 2021-11-14, 10x120L, EQMOD HEQ56, ZWO ASI071MC Pro_stacked.fitsM45, 2021-11-14, 10x120L, EQMOD HEQ56, ZWO ASI071MC Pro_stacked.fitsM45, 2021-11-14, 10x120L, EQMOD HEQ56, ZWO ASI071MC Pro_stacked.fitsM45, 2021-11-14, 10x120L, EQMOD HEQ56, ZWO ASI071MC Pro_stacked.fitsM45, 2021-11-14, 10x120L, EQMOD HEQ56, ZWO ASI071MC Pro_stacked.fitsM45, 2021-11-13, 10x120L, EQMOD HEQ56, ZWO ASI071MC Pro_stacked.fits

[Paul M]

Ah, it looks like I lost a paragraph or 2 in my first post today; where I explained that DSS would only stitch 4 of the frames and struggled with one of those. Even when the individual frames where stacked first. I think it just uses pattern matching and not astrometric alignment. Then I returned to ASTAP which is what my method above is describing. 

 

 

[han59]

The image have a lot gradient do to over-correction of the flat. The bias /flat darks where missing during stacking and probably causing this effect.  I tried to correct that with an artificial flat and some other tools with limited success. The M45 tiles are difficult  to correct.

Anyhow I used them to test the tool "ring equalise"  in ASTAP, tab pixel math 1 and made some improvements for next ASTAP edition. Below the result before I gave up. It  is maybe a little better but still not good. For a good result "flat darks"are essential.

ICE in general will produce better result but I don't know how good it is with gradients.

Han

 

 

 

Edited November 16, 2021 by han59

[Paul M]

50 minutes ago, han59 said:

The image have a lot gradient do to over-correction of the flat. The bias /flat darks where missing during stacking and probably causing this effect.

I'm somewhat lazy, it has to be said!

I've only been taking light flats with this camera, which I've had very few opportunities to use since purchase.  

When I stacked in DSS the flat certainly looked over corrected but ASTAP, to my eye, made much better use of the master lat it produced.

Anyway, thanks for your effort Han, it's much better than my attempt. I could cool the camera again and take all the bias/dark/and flat frames as nothing has moved in the optical train, not that it matters for bias and dark flats! Then try again

Edited November 16, 2021 by Paul M