Drizzled Saturn

[asimov (John)]

Any of you guys tried Registax drizzling on undersampled data yet? These two were captured @ F20 (F10 SCT) & then drizzled 1.5X. Resized down a tad to suit the seeing for each.

Anyway just as a matter of interest, drizzling does work well under certain conditions.

[ballaratdragons (Ken)]

Well worth the effort AsiJohn

The surface looks like the planet is spinning at about a thousand RPM

Those surface colours are nice
Good job

[cybereye (Mario)]

I've been playing with Drizzle in both Registax6 and AV!2. Unfortunately the rubbishy data I've had to work with hasn't given me anything worth posting!!

Cheers,
Mario


PS I'm starting to prefer AV!2 over Registax6 - it seems to give me slightly better results, especially on poorer data.

[lepton3 (Ivan)]

I have used drizzle in R6 on lunar data to good effect, so long as the data is reasonably sharp and there are lots of align points. This is my default method, as it gives a nice large image from my small 640x480 DMK21 camera.

Tried once on AS!2 with some mediocre jupiter data, results was just a fuzzball.

That is a great Saturn result you have there.

-Ivan

[Quark (Trevor)]

Really nice results there Asi, beautiful.
Have never tried drizzling, ever.

Regards
Trevor

[asimov (John)]

Thanks guys. Obviously just playing around at this stage but gee, it certainly has a place in the big scheme of things I think..

[Shiraz (Ray)]

beautiful images John. Yeah, have used drizzle in the past when there is a chance that the image has been slightly undersampled, in an attempt to extract the last possible skerrick of detail - but the seeing has not been anywhere near good enough for that to be an issue for the last few months.
Thinking about it though, there might be a good argument for using undersampling to push up the framerate on Saturn to get around the seeing a bit and then using super-resolution techniques to get back to full res. Will give it a try - thanks. Regards Ray.

[asimov (John)]

Thanks Ray. The limit for the C11 & OSC camera right now is basically F20 although I have had some success with F30 & a 35% full histogram. I think really good seeing is a must @ F30 with my setup though.

Yes, with drizzling, good seeing data is almost mandatory it seems. I've tried a few in average seeing & just laughed & sent them on their way to the recycle bin.

[desler]

Lovely, both of them Asi.

Darren

[asimov (John)]

Hi Darren. Thanks mate!

[Shiraz (Ray)]

Hi Asi. as promised, I tried deliberate undersampling and restoration with drizzling with a mono camera under various conditions - looks useful in intermediate seeing.

First image is red channel data slightly undersampled at collection (~f20) and then 1.5 resampled prior to processing using interpolation. Second image is the same data with drizzle super-resolution (AS!2) to 1.5x. Third image is from a similar run but at the normal ~f28 and with slight scaling prior to processing (sorry about the angle variation). All three were sharpened in IRIS using the same deconvolution routines.
The main difference was that the undersampled data could be acquired at 60hz whereas the nomally sampled data was gathered at 30hz.
The drizzle result is higher res than simple resampling and both are considerably better than normal sampling in this set of conditions - the extra sampling speed helped cut through the seeing. Of course you don't get anything for nothing and the the drizzled data is noisier, but overall, there is still an advantage in undersampling. Will be using this method on dim targets from here on and will have to try it with more severe undersampling.

You are onto a good additional weapon in the processing stakes - be interesting to see what you get by pushing the envelope with your Bayer camera. Not altogether sure what the Bayer matrix does to super-res processing. Thanks for the thread.
Regards Ray