16 replies to this topic

[james7ca]

This is a single-frame capture using a Nikon Z8 and a Celestron 9.25" EdgeHD with a 0.7X reducer. There are additional capture details in the image captions.

 

I also have a burst capture sequence that I did with this same equipment that consists of 1751 frames but that's going to take some time to process.

 

The so-called Lunar X and Lunar V are visible along the terminator.

 

Image processing with Photoshop 2024 and PixInsight.

Attached Thumbnails

• 

[Tom Glenn]

James, a very nice image.  How do you like the Z8 with the C9.25 EdgeHD for lunar imaging?  Was there a particular advantage to using the reducer here?  It appears that the full disk of the Moon should fit even without using a reducer.  

[scottinash]

Lovely image, James!   X and V are definitely showing strong!  

Edited by scottinash, 16 April 2024 - 01:24 PM.

[Borodog]

Excellent image, James.

[james7ca]

Tom, Scott, and Mike, thanks for the notice. Also, I appreciate the likes.

 

As for the Z8 and lunar imaging, it's a pretty good combination since its stacked CMOS sensor allows full-frame, RAW burst rates of up to 20fps (using the camera's internal memory buffer) and it has a fully electronic shutter. During last night's session I averaged about 14fps over two minutes of capture but I was using a 1:1 format (5504x5504 pixels) rather than full frame (8256x5504). In any case, I'm not using the fastest CFExpress card that you can buy and some people report being able to do sustained 15 to 20fps even in full frame (although I think the 20fps reports are highly suspect).

 

In terms of image quality, the stacked sensors don't seem to offer quite the same merit that you can get from a similar-sized, similar-generation CMOS counterpart but I doubt that matters much for lunar imaging. For DSOs I'm still not certain whether or not this camera is fully usable since it seems to suffer from the same colored, concentric ring problem as do all recent Nikon cameras. Under some conditions the latter doesn't seem to be a significant issue, in other situations (subject, sky brightness, exposure, vignetting) it can cause problems (or so it seems from my limited use).

 

One aspect on the Z8 that I still want to investigate is its ability to do pixel-shift captures. I suspect that it might be useful for extreme wide-field work with short-focus camera lenses since the 4.34um pixels on the Z8 don't really sample stars that well under those conditions (with a good lens). Basically you'd use the pixel-shifting to produce a guaranteed dither and then drizzle the subs. However, a 2X drizzle would produce a 180 megapixel output which is getting a little much. Unfortunately, I don't think the pixel shifting would offer any real benefits when imaging with a scope since at that kind of scale the seeing conditions would be moving the image around more than the pixel shifting itself. But it does offer an easy way to do accurate 1 to 2 pixel dithers on your captures. In any case, I'm waiting for the day that we can have a dedicated astro camera that has sensor shift capabilities built-in since it makes a lot more sense to move just the sensor than the entire mount and telescope during fine guiding. And, one would hope that sensor shift technology would be much cheaper than today's AO units.

 

One of the reasons I was using the reducer was that my EdgeHD was still setup for some work I did on earthshine a few days ago ( __HERE__ on CN) and in any case I wanted to see how it would perform on DSOs. I still need to get a so-called "wide" T-adapter since the M42 variety I have now vignettes pretty badly (on the full-frame Z8). However, if it is clear tonight I may switch back to the f/10 focus since the moon is getting a little bright for DSO work. I suspect that the results at f/10 will be even better than you might expect just from the change in the image scale since Celestron's 0.7X reducer isn't really known for providing top-notch image quality.

Edited by james7ca, 16 April 2024 - 11:17 PM.

[james7ca]

Since Tom was wondering about the Z8 at the f/10 focus of the EdgeHD below is a quick crop and processing of ONE FRAME at full image scale as captured on Tuesday night (the full frame included the entire moon). As before, I have a thousand plus additional frames to stack which will certainly result in a better image. In fact, I was focusing fairly cleanly on a group of small craters that are below the northern rim of the crater Ptolemaeus that aren't even visible in this singe frame, so a sharpened stack should show a lot more detail.

Attached Thumbnails

• 

[james7ca]

Well, I selected 64 random subs from Tuesday night's session and then use AutoStakkert! to combine the best 48 of those followed by a wavelet sharpening in Registax to produce the following image. This is based upon the same session where I processed a single frame as shown above and as expected a stack produced a much more detailed result. I actually have almost 1700 frames in the entire series and thus this trial is still less than 4% of my data. You can now see the small craters on the floor of Ptolemaeus that I mentioned earlier, the ones I used to check focus. Again, this is a small crop from the full-disk capture.

 

I wonder what a drizzle would look like that was based upon the full 1700 frame sequence. Probably couldn't get twice the image scale but it might reveal some additional detail. Next up (tomorrow possibly, if it is still clear) is to use a 2X barlow as that will get me closer to critical sampling with the Z8. However, I doubt that a barlow would cover the full frame of the sensor and of course at that increased scale I would have to do a mosaic to cover the entire disk of the moon.

Attached Thumbnails

• 

Edited by james7ca, 17 April 2024 - 04:21 AM.

[flt158]

Excellent images from you, James7ca. 

You got Rupes Recta extremely close to the lunar terminator. 

I love it!

 

Great imaging, Aubrey.  

[scottinash]

Very well done!  

[james7ca]

Aubrey and Scott, thanks for the comments.

 

I think the sharpening and tonal quality on 48-sub stack are probably less than ideal, but for such a "short" stack it's probably okay and I didn't want to spend a great amount of time on the processing since it was simply a test.

Edited by james7ca, 17 April 2024 - 07:21 AM.

[Borodog]

I’m excited to see a full stack at f/10, and interested to see a drizzle. How was the seeing during capture?

[james7ca]

The seeing conditions were pretty good since the moon was high in the sky and we were not having a Santa Ana condition (the latter is often the case when the skies are clear at night here in San Diego and the Santa Ana winds always bring very poor seeing conditions).

 

Any kind of drizzle is going to require some extra work since I don't think AutoStakkert! or my PC with 64GB of DRAM can drizzle images that are as big as output by the Nikon Z8. So, I'll have to crop the originals down to a reasonable size before processing. The Z8 captures 46 megapixel images when in full-frame mode but I used a 1:1 format for these lunar images and that only has 30 mega pixels but with the 3X drizzle in AutoStakkert! that would mean a drizzled result of 9 x 30 or 270 megapixels. Converting that to 16-bit RGB would mean a file size of something like 1.6GB.

[james7ca]

Here is my first attempt at using all of the subs that I captured using the Nikon Z8 and the EdgeHD at its f/10 focus. The original frame covered the entire moon, but here I cropped an area around the crater Ptolemaeus and did a 1.5X drizzle so that this can be compared directly with post #7 where I stacked only 48 subs without a drizzle.

 

The image seems to have held up pretty well with the drizzle and what you see below is 1.5 times the original capture scale (as shown in post #7). Plus, I allowed AutoStakkert! to select the best 640 frames out of the 1685 subs that I captured in burst mode on the Nikon Z8.

 

I think this confirms that a drizzle should work on future captures with the EdgeHD and that gets me a little closer to critical sampling with the Z8's 4.34um pixels. Thus, the effective pixel size with the drizzle is 4.34um / 1.5 or 2.89um. Not quite in the same league as Sony's IMX183 with its 2.4um pixels and of course the Z8 is a color camera where capturing in mono has certain advantage on lunar imaging.

 

I've gotten better images of this area before when using the EdgeHD, but I think my processing with the Z8 can still be improved and it is nice to be able to capture the entire moon in a single frame (something that I've never been able to do before at this image scale).

 

I also suspect that I might be able to use the Z8's pixel shift ability to dither a bit more uniformly since for this capture I relied on the drift in the tracking and that didn't give equal step coverage/magnitudes in both X and Y. 

Attached Thumbnails

• 

Edited by james7ca, 22 April 2024 - 07:52 AM.

[james7ca]

I measured a few craters using the LROC Quickmap and its geodesic (circle) tool and I came up with a minimum resolvable crater diameter in my drizzled image of around 1.4km which works out to 0.75 arc seconds given the then distance between the earth and moon. There are craters that are somewhat smaller than this that are what I'd call detectable, but I only used craters that seemed to have a clear dark to bright rim in my image. Dawes' limit for the 9.25" EdgeHD is 0.5 arc seconds so I suspect that there is still some resolution to be had with either smaller pixels or a longer effective focal length. In fact, I've resolved craters as small as 0.8km with the EdgeHD and that worked out to just under 0.5 arc seconds.

 

That said, does anyone have a reference to a detailed map of the floor of the crater Ptolemaeus showing crater sizes down to about 1km? I seem to remember seeing such a map on the internet but I did a search and couldn't find any match.

Edited by james7ca, 22 April 2024 - 09:10 PM.

[Tom Glenn]

That said, does anyone have a reference to a detailed map of the floor of the crater Ptolemaeus showing crater sizes down to about 1km? I seem to remember seeing such a map on the internet but I did a search and couldn't find any match.

I also recall such a map, although the only place I could find it was in a copy of The Cambridge Photographic Moon Atlas, Chu et al. (2012).  I took a photo of the map and reproduced it below.  The credit in the caption says the measurements are the work of M. Clark using data from the Lunar Orbiter Atlas.  I believe this is an error, and should read the Consolidated Lunar Atlas, because this is clearly an Earth-based image and not part of the Lunar Orbiter series.  I believe M. Clark is Mardi Clark, who used to publish a website (that appears to no longer exist) and post here on CN many years ago.  

 

The quality of the reproduction is not great (by today's standards) and does not include values for many of the smaller craters.  Additionally, I don't know the methods used to make the measurements, and a quick check with the LRO Quickmap shows that although everything is reasonably accurate, I would personally retake all measurements using LROC data if you want the highest accuracy.  Several of the listed values differ by at least several tenths of a kilometer, although this also is completely dependent upon where one takes the measurements from.  

 

[james7ca]

Tom, I agree that some of the measurements in the resolution chart seem a little suspect. However, I found three craters that were labeled at 1.3km that seem to be resolved in my drizzled image (and those were the "smaller" 1.3km craters, since there are a few at that size that seem larger than 1.3km in both my image and the map). The 1.3km size is consistent with the measurements I took on the LRO data although once you get to that size even a 100m change can be significant. But, I'd guess that the drizzled image may have gotten down to 0.7 arc seconds (i.e. a bit better than I stated earlier). I may try a 3X drizzle on the data although I seriously doubt that will reveal any more detail.

 

Thanks for the help in finding this resolution map, it does look vaguely familiar.

Edited by james7ca, 23 April 2024 - 01:10 AM.

[james7ca]

I found a thread on CN that has several crater maps with one of Ptolemaeus that has measurements down to and even below 1km. Here is that link:

 

  https://www.cloudyni.../#entry11568923

 

In my image I can see the only two craters that are marked as being 1.3km in size. Thus, this is another confirmation on that limit as recorded in the drizzled image from the Nikon Z8.