Here is what we believe is a composite image of Uranus with four of its moons. The outermost and brightest is Titania, next, close in to the planet, is Miranda. On the opposite side are Ariel (closest) followed by Umbriel. Positions derived from the Sky & Telescope javascipt 'Uranian Moons' .(http://www.skyandtelescope.com/wp-content/observing-tools/uranus_moons/uranian.html
Very difficult to image the moons (around 12-13th magnitude at best) without overexposing the planet into a bloated bright blob. As such this is a composite of two separate photos: one of the moons and one of only the planet itself.
Imaging train and specs:
8" refractor @ f/12. QHY5L-II mono. Seeing decent, 7/10, degrading to variable - it began to get cloudy during the full disc imaging run. The date and time: December 16th @ 00:14 UT.
For the moons, the QHY gain was 88% and the exposure was 5 seconds. The camera was set for 'Long Exposure Mode'. The moon image is a stack of 65 of these frames, with a dark frame applied.
For the planet, the QHY gain was 27.8 and the exposure was 100 ms (0.1 seconds). A stack of roughly 1000 frames.
Note that during capture in EZPlanetary, the histogram for the 'moons' exposures is a blown out spike where the planet sits and that the moon data barely, if at all, spikes out of the noise level. This is a result of the very high gain and long exposure time. During the 'planet' exposure, the histogram is much more representative of the kind seen during normal (i.e. Mars, Jupiter, or Saturn) planetary imaging.
The moons will stack out in Registax, and further processing with the wavelets will sharpen their discs. Photoshop does the rest in creating the composite.
It is interesting to note that WinJupos provided a close, but not exact match to the Sky & Telescope javascipt. However, I am one version out of date on WinJupos, and the S & T script is expiring on 31 December after being implemented on 1 December 2008. I wonder if the S & T script is getting inaccurate?