Problem with Planning Mosaics

[twade]

To all,

I'm hoping someone on this forum has developed a means of properly framing each shot in the field.  I"m having a terrible time getting each frame to line up properly.  There is always considerable rotation between each frame which ruins the mosaic.  In fact, the centering of each frame gets skewed as well.  I use TheSky to create the framing with a 15 degree overlap between each frame, but apparently its underlying algorithms are too simple (i.e. it views the sky as flat).  This simple model does not work when using focal lengths less than 300mm.  For example, here are two 2-frame mosaics I created over the August New Moon period.  The North America nebula was a four frame mosaic, but it is hopeless to combine all four frames since the rotation is so bad there's very little overlap.  The rotation is not due to polar alignment.  It is entirely due to the sky being a sphere.

Using a 75mm lens:


Using a 165mm lens:


As you can see, none of the frames line up.  I need a means to calculate the "exact" rotational angle between each frame which should also compensate for the center RA/Dec coordinate so they line up edge to edge with the proper overlap.  Does anybody know of any software that will do this?  Perhaps, one of you has developed a procedure that does the same?  Are there any tools within PixInsight that may add in mosaic planning?  Any help would be greatly appreciated. 

My two nights of imaging were a complete waste of time since there is very little overlap in the resulting mosaics. 

Thanks,

Wade

[twade]

To all,

I'm going to try an experiment during the Full Moon window.  I'm going to keep the position angle at 0 degrees.  Hopefully, this will eliminate any rotation due to frame to frame Declination differences (i.e. each horizontal frame should differ in RA only).  I'll keep you posted.  Any other suggestions would be greatly appreciated.

Wade   

[ruediger]

Hello Wade,

you can try SkyChart (Cartes du Ciel) for mosaic planning. It allows to place as many camera panels as you need and save their position to a text file.

Rüdiger

[blackdragon72]

Good to learn this feature of Ciel Skychart.

[twade]

Rüdiger,

Thanks for the suggestion. 

Are there any tutorials on the use of this program, specifically generating these charts?  I'll download it tonight and begin experimenting.

Wade

[JoLo]

If by chance you have an Atlas and use EQMod, there is a really nice mosaic planning module.  It works with EQMod but is a stand alone and needs to be downloaded separately.

Joe

[mads0100]

Sequence Generator Pro has a 'Framing and Mosaic' Tool.  It takes data from SAMBA and basically tweaks it to your FOV so you can draw boxes around what you want to image.  It then figures out, with overlap, the coordinates for each frame.

Works pretty good.

[gvanhau]

Hello
Before planning the mosaic, you have to know your camera rotation angle and input this in TheSky.
In order to know your camera rotation angle, you can take one image and solve it.

The other way to go is to plan your mosaic first, (here you adjust the best angle for the object) and then in the field, you rotate your camera until you match the planned angle.

If the planned angle does not match your camera angle, there is no way your mosaic will match the planned one.

Geert

[twade]

Geert,

I believe you nailed it.  I plan my mosaics at home with the appropriate PA, but I can't quite get them exactly out in the field.  From now on, I'll plate solve a test image out in the field, determine my PA, and adjust the mosaic accordingly.  Now, I just need some clear skies.

Thanks,

Wade

[Greeps]

We had similar problems but not to the extent you posted.
We set up our mosaics at PA=0. Then I calibrated the rotator with plate solves.  This helped a lot although not completely.  We use an overlap of 8 degrees.

When a mosaic is shot, each panel in a vertical column of tiles will be similarly aligned, the adjacent column will be slightly tilted, the next column over will be tilted more.  If you number an 3x3 panel mosaic like a tic tac toe grid with 1 being at the upper left then 2 the upper center, etc., then panels 3, 6, and 9 would be identically aligned, almost perfectly. Panels 2, 5, and 8 would be rotated relative to the first column panels, but would be rotated identically to each other. 1, 4, and 7 would be the most rotated relative to the panels in the first column but, again, those 3 panels would all show the same degree of rotation.

Careful calibration of the camera and rotator to PA=0 with plate solves allowed us to get a 9 panel mosaic with a 500mm Tak FSQ106 using an 8 degree overlap. Our target was close to the celestial equator.  I suspect the rotation problem gets worse as a target is selected that is closer to the pole but we have not tested that.  I don't know why the problem exists. The mosaic as displayed in TSX looks perfect but the resulting panels in the light frames don't agree. We exported the layout straight from TSX so there was no opportunity for us to introduce the error.

I hope this helps. I wish I had a fix to offer you.

Best,
David

[CharlesW]

This would be an interesting question to post over on the Bisque support forum. I'd be interested to see what their solution is.

[Greeps]

Good suggestion.

I have just posted a detailed message in the Bisque Yahoo group. Perhaps someone there has a solution. I have not posted on the Bisque forum.

David

[mads0100]

Check out Sequence Generator Pro. They've already figured out the solution.