New Feature : Moon Mosaic Planner

[Borodog]

More alignment failures again last night. The scope and camera were 2980mm focal length and the ASI183MC.

Automatic easily fit the limb of the Moon this time, and after slewing a few times and making 3 sets of measurements, it declared the something was wrong with the alignment and to try again.

I tried manual but again could not eyeball the rotation accurately enough that the bottom of the mosaic was not completely out of frame when I slewed there to test.

I ended up just trying to shoot the mosaic manually.

[admin]

Hi,

any chance of the log from the failed alignment run? It will probably contain enough info for me to at least understand the problem.

cheers,

Robin

[Borodog]

Thanks Robin. I'll send it to you later today via email.

FYI I missed a spot in the manual mosaic:

https://cdn.astrobin.com/images/93441/2 ... 467d82.png

[admin]

Oh no, such a tiny patch too! I'd have been in there with the clone brush in photoshop/affinity photo I think

cheers,

Robin

[admin]

Just a quick update.

The latest version - 4.1.12489 - adds a new option to try to measure drift when performing an automatic mosaic align. If this option is selected, SharpCap will monitor the image for 20s at the beginning of the auto align procedure and try to work out if the image is drifting (perhaps as a result of polar alignment not being quite right). If a significant drift is detected then the measured drift rate will be used throughout the rest of the mosaic process to try to ensure that the panel captures happen where planned and are not out-of-place due to drift.

If you see a measured drift of zero in either direction, it means that (most likely) there was no statistically significant drift during the measurement process. This can be affected by the amount of instability in the image, so if there is lots of instability due to poor seeing, it will be less likely to spot drift as being significant.

I'd be interested to hear how it works out - in particular, the 20s measurement period might turn out to be too short, or need to be adjustable

cheers,

Robin

[Borodog]

Hmm. I'm intrigued but to be honest also skeptical. I am doubtful that 20s is long enough because it seems like drift will be dominated by periodic error. In fact, I don't see how this can work well if the measurement period is shorter than the RA period. Or am I missing something?

[admin]

Hi,

I think the dominant error is likely to be declination drift caused by poor polar alignment, at least that's what seems to be coming up in a number of real life cases where I've been trying to help people get their mosiacs to run more accurately. RA periodic error might be an issue that could need dealing with by excluding RA from the correction (although it all rather depends on how long your periodic error period is compared to the time taken to capture a mosaic I guess)

cheers,

Robin

[Borodog]

Ok; it makes sense that if you are correcting you have to evaluate the relationship between the screen/camera axes and the mount axes, as in Feature Tracking. I seem to recall that you were reticent to do that before for the mosaic planner, although perhaps I am mistaken. In any event that would easily allow you to only correct in DEC. Although it seems as if this functionality is redundant to Feature Tracking.

[admin]

Hi,

you can't use feature tracking during a mosaic capture, as the movement to get to different panels will upset the feature tracking greatly!

I'm now getting to the stage where I have a more general set of code handling the orientation/scaling requirements between the camera axes and the mount co-ordinate axes, so some things that were hard to do before are now getting easier to achieve. Right now, the mosaic tools don't quite co-operate with that code, since they work via camera -> ecliptic -> equatorial co-ordinates (ecliptic co-ordinates let you get the orientation of the sun/moon sorted nicely), so they are a bit more complex than the other tools which are just camera <=> equatorial co-ordinates.

cheers,

Robin

[Borodog]

I think you would want to suspend Feature Tracking for a move to the next panel and then restart it. Or, base the mosaic tool on image registration and Feature Tracking and have SharpCap use Feature Tracking just drive the scope to the next panel, ensuring the alignment is correct via image registration (e.g. features previously on the right are now on the left). I think this method is more robust than blind slews, especially for critically sampled imaging with a long focal length, where the FOV seems like it is always going to be too small for accurate calibration using any of the mosaic calibration methods. And having Feature Tracking compatible with the mosaic tool would also fix drift, of course.

In the meantime the next time out I'm going to try to use the ghosted image alignment and do the mosaic manually, since it's compatible with feature tracking. Setup up for a panel, and before starting capture, grab a snapshot for use in the GIA. While building the mosaic you can build a library of snapshots to quickly check the left/right and up/down alignment with the previous panels. Very tedious, but at least doable. I just don't see any of the current mosaic calibration methods ever working with my 1100 EdgeHD and either of my 2 micron cameras, and the drift is just a deal breaker.