combining multiple sources at different image scales in planetary photography

andrea tasselli:

Tony Gondola:
I doubt that you could get the accuracy you're looking for with manual rotation and scalin

*You'd be wrong.

andrea tasselli:
In IRIS open the image that needs to be the reference one, namely the luminance. Let's assume the both the RGB and the L image are perfectly flattened and they are taken at the same time (or, equivalently, with the same phase and libration angle) which, depending on the source/target image scale might restrict the lag between source and target image. Type the command

>CIRCLE threshold  

from the console using a guess threshold value (or hover with the mouse pointer on the moon's limb) for the limb brightness (intensity), drawing a box around the Moon's image. Check that the circle fully encompasses the whole Moon limb just at the edge of it and note down the radius and the center coordinates. Do the same with a new instance of IRIS for the target RGB image. Also, keep in mind that the distribution of brightness might not be perfectly circular but as long as they correlate each other closely that shouldn't matter. The ratio between the two radii is the scaling factor to be applied next to the the target image, with the command:

>SCALE 3 scaling_factor scaling_factor

You might to have to juggle a bit with trial values here to get a perfect match and it is important to have the two images in the same orientation.

Then register the two images using the planetary registration command:

>PREGISTER [IN] [OUT] [NUMBER]

where SIZE is a power of 2 number setting the correlation window size, so let' say 1024, IN and OUT are the generic name of the two images, with the first being the reference image (L frame) and NUMBER is 2. Now, if everything goes well the two images are aligned and with the same size. The last bit is to rotate the RGB image as to match the same features on L image, which shall be accomplished by using the ROT command using the results at step one.

>ROT XY YC Angle

where XC and YC are the coordinates of the center of the RGB image of the Moon (NOT the coordinates of the center of the frame!). Angle is obviously the angle you need to rotate the RGB  image to match the L image. It is helpful to use just a single channel (the G one is best) to carry out the trial rotations so that you can use the SUB command to see how well you are doing (or from the Processing menu the command subtract).

Expect a significant amount of fiddling here and there but you should get there in the end.

Tony Gondola:
That wouldn't be possible in the present version of Siril. Alignment would be easy in any photo editor but I doubt that you could get the accuracy you're looking for with manual rotation and scaling. Why not just use the 10" with OSC?