Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Crazy About Color?

  • Please log in to reply
27 replies to this topic

#1 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 08:32 AM

I know the moon doesn’t have much in the way of color, but the landscape photographer in me really loves color.  While I thoroughly enjoy high resolution b&w lunar images, I’d like to add some color into the mix.  Finding color images by Bartosz Wojczyński earlier this year, and also one by Ignacio Diaz Bobillo (you can see his image at http://www.pampaskies.com/gallery3/Moon/Luna_Cenital_Color) really put the bug in me to try to get ‘high resolution’ lunar images in color.

 

Now that I’ve been researching the topic more, I have of course discovered some of the major hurdles associated with lunar photography, particularly in color.  I’ve been looking into a color CMOS camera (QHY 183C) to be paired with a Celestron C8 EdgeHD telescope.  Not wanting to go into a big mono/color wheel versus color camera debate, I do understand that the Bayer matrix will reduce the effective resolution of the color camera versus a mono camera of the same model.  I believe this can be abated somewhat by oversampling of the image, say by shooting at resolutions of .15 to .1 arcseconds and then resizing the image at an appropriate resolution.

 

The other challenge I recently learned (or at least was reminded from my high school physics class) was light refraction through the earth’s atmosphere for different wavelengths of light.  Granted, that is an issue whether I use a color camera or mono camera.  I believe I can address this issue with the use of an ADC, such as one offered by Pierro Astro, as well as by using FireCapture to align the color channels.  I’ve seen examples of shots of Saturn using this setup, but the angular size of Saturn is small compared to the moon.

 

Does anyone know if the ADC works with a large size senor like the QHY 183C (13.3mm x 8.9mm)?  Can I maintain a flat field with the ADC?

 

I realized in my looking at higher resolution lunar images (I’ll informally say ‘high resolution’ is at 0.25 arcseconds/pixel or less), I’ve almost seen none in color.  Seeing the truly outstanding and real high resolution B&W images taken by people on this forum, I now recognize most of the color images posted by Bartosz Wojczyński, as good as they are, are not high resolution.  Bartosz has only one high resolution image and it was taken in bicolor with a mono camera.  Ignacio’s image (with the link above), was taken with a DSLR and processed with separate channels and recombined in PixInsight, and also is a bit shy of high resolution.  I’ve not found a single color image (doesn’t mean it doesn’t exist, I just haven’t seen it with hours of looking) of the moon in high resolution that was taken with a color camera and used an ADC.  Does anyone have an example of an image with a resolution of at least 0.25 arcseconds/pixel with this setup?

 

I’ve started to realize that if I don’t see an image or process readily posted online, there’s generally a good reason for it.  Is high resolution color photography of the moon a pipe dream?

 



#2 Bart Declercq

Bart Declercq

    Viking 1

  • *****
  • Posts: 676
  • Joined: 21 Jan 2007
  • Loc: Haaltert, Belgium

Posted 07 December 2017 - 09:01 AM

One thing you need to take into account is that an ADC doesn't work very well with short F-ratio's (ZWO advises using it at F/20 or higher) - which clashes with using a tiny-pixel sensor (at 2.4µm the pixels of the QHY183C will reach a telescope's diffraction limit at an F-ratio of about 12, well below optimum for an ADC.



#3 PETER DREW

PETER DREW

    Ranger 4

  • -----
  • Posts: 377
  • Joined: 31 May 2017

Posted 07 December 2017 - 09:04 AM

Lunar colour is too subtle to be seen visually easily even through a telescope, it maybe that at the higher magnification used for high resolution photography it's still too subtle for a camera? 



#4 happylimpet

happylimpet

    Surveyor 1

  • *****
  • Posts: 1990
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 07 December 2017 - 09:25 AM

As Bart says, an ADC will require a high f-ratio, though 20 isnt necessary, even 10 is OK according to some sources. However, if the Moon is high (say above 45deg, which you could select) there should be no real need for an ADC.

 

Colour cameras dont reduce resolution if you use AutoStakkert to stack (which you would anyway, as its the best), as it uses natural dithering to fill in the gaps without any interpolation.

 

Theres no reason why a standard colour CMOS camera (like the ASIs) cant be used to make high res lunar colour images.

 

Heres a colour image I took with my Canon 500D. I just stacked a few hundred frames to get very high signal/noise, and stretched the colour in photoshop. (Not high-res, I know!)

 

10887666_10152536989510689_.JPG

 

The only lunar colour Ive detected visually is the orange area around Aristarchus.


Edited by happylimpet, 07 December 2017 - 09:26 AM.

  • RedLionNJ, xiando and aeroman4907 like this

#5 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 09:29 AM

Bart, I was thinking of imaging at f/20 (possibly f/25) with this setup, particularly to address the Bayer matrix issue anyway.  I'm also looking to do large lunar mosaics, so I'm looking for both a small pixel size and relatively large sensor size so I can keep my panel count down to create the mosaics.  I figure at f/20 with that setup I'll need 24 panels to create a full moon image.  Seeing is a challenge here, which I know trumps all, so I'm trying to be able to get as much imaging done in a 'short' time as possible before the seeing changes dramatically.  I figure 24 panels is at least an hour's worth of good imaging.



#6 beggarly

beggarly

    Ranger 4

  • -----
  • Posts: 372
  • Joined: 12 Jun 2015
  • Loc: Belgium

Posted 07 December 2017 - 09:38 AM

Some examples. Unfortunately there's not much info.

http://www.astronomi...on_color_en.htm



#7 happylimpet

happylimpet

    Surveyor 1

  • *****
  • Posts: 1990
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 07 December 2017 - 09:51 AM

You may also want to consider the following approach, if you want to do some petrology....

 

http://www.rkblog.rk...and-age-colors/


  • Vedran likes this

#8 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 10:15 AM

Thanks guys.  Perhaps I will be blazing a newer trail (or spectacularly failing in the attempt).  I of course believe B&W moon images are great and worthwhile taking, but I don't feel I'll be adding anything meaningful that hasn't already been done by more experience people with much better equipment ( I don't have $50,000+ to drop on a Planewave 24 inch CDK telescope to try and improve on the mono images that already exist in plenty).

 

Regarding color, I'm going for a somewhat 'subtle' approach, similar to the attached image.  I'll be putting much more effort into getting the color 'just right', but at least this is a personal example.

 

Very Final RGB.jpg


  • John Miele likes this

#9 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 01:12 PM

Thanks Happylimpet, you make a good point about the altitude of the moon in the sky.  As I'll only be seriously imaging under close to the best of circumstances, I'll of course be imaging the moon high in the sky.  Any of the moon phases that require imaging lower in the sky are going to have a lot more turbulence anyway, so varying color diffraction will likely be one of my least problems.  This saves me some money, learning curve, and hassle for now in getting the ADC.

 

I'll probably still get an ADC for planetary work, but that will be more pedestrian and for pleasure of having my own image.  An 8" aperture is a bit small for obtaining any real impressive imaging of any of the planets.  Even with Mars this coming year getting an image of Mars 200 pixels across that looks good would be a massive win, but having learned what it takes to upsize and work with 800 pixel wide images of the moon has taught me there is only so much you can do.  A 200 pixel image of Mars is only 1-1/3" printed out at a lower resolution of 150 dpi.



#10 JMP

JMP

    Viking 1

  • *****
  • Posts: 561
  • Joined: 31 Oct 2005

Posted 07 December 2017 - 02:11 PM

I use a C8 and a color camera. I think color adds a bit of "life" to the image even though "colorized" images strike me as a bit phony. There is no limit to the amount of money you can spend, and it's easy to become OCD about how close to perfect is good enough. When you look at the cost of a good full frame DSLR it's more than enough buy that C8-HD.

 

I'll attach a nice close-up of Plato, shot with a non-edge C8, and an older ASI-120mc color camera. This is cropped from the 1280 by 960 original so you can see the full resolution on the cloudynights page. I like to shoot close-ups at about 6 pixels per arcsecond, this seems to work well with a color cam and a C8.

 

Jeff Phillips

 

Plato(Crop).jpg

 

And a similar crop of Copernicus:

 

Copernicus(Crop).jpg


  • happylimpet and aeroman4907 like this

#11 Wildetelescope

Wildetelescope

    Viking 1

  • -----
  • Posts: 783
  • Joined: 12 Feb 2015
  • Loc: Maryland

Posted 07 December 2017 - 02:26 PM

Thanks Happylimpet, you make a good point about the altitude of the moon in the sky.  As I'll only be seriously imaging under close to the best of circumstances, I'll of course be imaging the moon high in the sky.  Any of the moon phases that require imaging lower in the sky are going to have a lot more turbulence anyway, so varying color diffraction will likely be one of my least problems.  This saves me some money, learning curve, and hassle for now in getting the ADC.

 

I'll probably still get an ADC for planetary work, but that will be more pedestrian and for pleasure of having my own image.  An 8" aperture is a bit small for obtaining any real impressive imaging of any of the planets.  Even with Mars this coming year getting an image of Mars 200 pixels across that looks good would be a massive win, but having learned what it takes to upsize and work with 800 pixel wide images of the moon has taught me there is only so much you can do.  A 200 pixel image of Mars is only 1-1/3" printed out at a lower resolution of 150 dpi.

That depends a lot on your seeing conditions.  I have seen some REALLY nice Mars, Saturn and Jupiter images from a C8 from a friend down in Florida, where the sky is very stable.  To be sure, we have been trying to convince him that he REALLY needs a C11 or C14, given his skill, but his C8 images are nothing to sneeze at.   All depends on your expectations.  

 

Cheers!

 

JMD


  • happylimpet likes this

#12 Wildetelescope

Wildetelescope

    Viking 1

  • -----
  • Posts: 783
  • Joined: 12 Feb 2015
  • Loc: Maryland

Posted 07 December 2017 - 02:28 PM

This is an interesting question.  I am curious, does anyone know what we SHOULD expect from color on the moon, even at low mag images?   I typically see some green and brown, particularly in the Mare, but I have always wondered what a realistic representation of the color would be. 

 

Cheers!!

 

JMD



#13 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 02:35 PM

Thanks Jeff!  Super good and resolute images for a conventional C8!  That's right around the resolutions I am hoping for and an excellent example of what is possible.  Good to see the color sensor can work well.  I'll clearly do B&W as well, and certainly understand your perspective on 'colorized' images.



#14 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 03:01 PM

JMD, the answer to the colors on the moon are more interesting than I thought.  Check out this link and a copy of the image is attached.

 

https://www.nasa.gov/mission_pages/LRO/multimedia/lroimages/lroc20100910_color_moon.html

 

True Colors of the Moon.jpg


Edited by aeroman4907, 07 December 2017 - 03:01 PM.


#15 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 03:19 PM

JMD, congratulations, now you got me wayyyyy more interested in this color thing.  Here's a super rough and fast color balancing to get it somewhat closer to the NASA image.

 

Actual Moon Color Processing.jpg



#16 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 07:18 PM

Okay, put a bit more effort into performing a reproduceable process for color matching for the moon.  Below is my effort.  Will follow with a second post with the NASA color map of the moon.  I think the colors are quite close.

 

Very Final RGB 1200 Color.jpg


  • Wildetelescope likes this

#17 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 07 December 2017 - 07:22 PM

Here is the NASA color map.  I adjusted the overall levels slighlty because the had the Mare so dark, but not color or saturation adjustments.

 

CN real moon color.jpg

 

 



#18 Wildetelescope

Wildetelescope

    Viking 1

  • -----
  • Posts: 783
  • Joined: 12 Feb 2015
  • Loc: Maryland

Posted 07 December 2017 - 11:15 PM

JMD, the answer to the colors on the moon are more interesting than I thought.  Check out this link and a copy of the image is attached.

 

https://www.nasa.gov/mission_pages/LRO/multimedia/lroimages/lroc20100910_color_moon.html

 

attachicon.gifTrue Colors of the Moon.jpg

So that is cool!  The NASA image is a great reference.  Now I have to go back and reprocess all my moon images:-)!  

 

JMD



#19 Wildetelescope

Wildetelescope

    Viking 1

  • -----
  • Posts: 783
  • Joined: 12 Feb 2015
  • Loc: Maryland

Posted 08 December 2017 - 12:27 AM

full mosaic.stretchedjpg
 
This is typically what I see.  I am using a one shot color camera, a Mallincam SSIc.   Supposedly has a 16 bit color depth.  This was shot in prime focus with an f8 127 mm Triplet.  4 images have been stitched together in FIJI.  Bit of post processing in Astra image and FIJI, including Deconvolution and wavelet for sharpening up and basic stretching of the image.  Not much done with the color per se.   You can see the contrast within the Mare and a bit of variation of green an brown.   I am sure there is more I could do with the color, and I will probably play around with it.  
 
Cheers!
 
JMD

  • aeroman4907 likes this

#20 Bart Declercq

Bart Declercq

    Viking 1

  • *****
  • Posts: 676
  • Joined: 21 Jan 2007
  • Loc: Haaltert, Belgium

Posted 08 December 2017 - 03:42 AM

Bart, I was thinking of imaging at f/20 (possibly f/25) with this setup, particularly to address the Bayer matrix issue anyway.  I'm also looking to do large lunar mosaics, so I'm looking for both a small pixel size and relatively large sensor size so I can keep my panel count down to create the mosaics.  I figure at f/20 with that setup I'll need 24 panels to create a full moon image.  Seeing is a challenge here, which I know trumps all, so I'm trying to be able to get as much imaging done in a 'short' time as possible before the seeing changes dramatically.  I figure 24 panels is at least an hour's worth of good imaging.

The Bayer matrix is far less an issue than many people seem to think, particularly when using Autostakkert's intelligent debayering (which doesn't debayer individual frames but just fills the pixels with the right colour and uses the small movements between frames to cover all pixels in the stack with all colours, as long as your stack has a few hundred frames you essentially don't lose resolution to the matrix.

 

I've done quite a few monochrome mosaics of the entire moon (including a completely insane 170Mpixel mosaic way back in 2010 which took me three hours of recording 270 video's and some 30-40 hours of processing - see https://flic.kr/p/p3bAVT ) - I've now done the same using an ASI1600 where it took me ~30 video's taken in ~25 minutes - https://flic.kr/p/DX4p7h so yay for technological advances :)

 

The QHY can do 15fps@20Mpix and in my experience you need to record between  500-2000 frames per video (depending on the seeing mostly) to get a quality end results, so at 2000 frames including the time to reframe, you'll need about 2.5 minutes per frame, which comes to exactly 1 hour (or ~20 minutes when doing 500 frames) - you'll need sufficient storage too, 25 raw 2000 frame video's at 8bit per pixel and 20 Megapixel is about a terabyte of data and it'll have to be on SSD since the througput to record at 15fps is about 300Megabytes/second, which can only be done on SSD!


  • RedLionNJ, happylimpet and aeroman4907 like this

#21 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 08 December 2017 - 08:19 AM

Beautiful image JMD.  It will be interesting to see how processing video works.  There are a number of adjustments I can do in single shot RAW files in conversion to a 16 bit image file (nothing crazy, but to get some level of color to show up).  After images are stacked, I have a work around to convert the image file to a 64 bit file XIFS in PixInsight.  All my color saturation adjustments are performed in the 64 bit space.  Once I'm done with that, I save as a 16 bit TIFF and then make the color corrections in PS.  The PI purists would say do those color corrections in PI, but PS is so much more intuitive and powerful in this regard.  If I could do it in the 64 bit space, I would, but all that extra effort is not worth the return.  As it is, it takes some time in PS to get it just right.

 

Bart, those are amazing images!  What scopes were you using?  Regarding the computer, I have a very high end laptop with a good sized SSD for my business.  The SSD is not 1TB though, so I'll have to transfer files during about half way through a full moon capture to one of the mechanical drives and then finish imaging.  Now I've got to be disciplined today and get more work done on my real business that currently pays the bills.  I can get a little caught up in this stuff!



#22 Bart Declercq

Bart Declercq

    Viking 1

  • *****
  • Posts: 676
  • Joined: 21 Jan 2007
  • Loc: Haaltert, Belgium

Posted 08 December 2017 - 09:33 AM

Beautiful image JMD.  It will be interesting to see how processing video works.  There are a number of adjustments I can do in single shot RAW files in conversion to a 16 bit image file (nothing crazy, but to get some level of color to show up).  After images are stacked, I have a work around to convert the image file to a 64 bit file XIFS in PixInsight.  All my color saturation adjustments are performed in the 64 bit space.  Once I'm done with that, I save as a 16 bit TIFF and then make the color corrections in PS.  The PI purists would say do those color corrections in PI, but PS is so much more intuitive and powerful in this regard.  If I could do it in the 64 bit space, I would, but all that extra effort is not worth the return.  As it is, it takes some time in PS to get it just right.

 

Bart, those are amazing images!  What scopes were you using?  Regarding the computer, I have a very high end laptop with a good sized SSD for my business.  The SSD is not 1TB though, so I'll have to transfer files during about half way through a full moon capture to one of the mechanical drives and then finish imaging.  Now I've got to be disciplined today and get more work done on my real business that currently pays the bills.  I can get a little caught up in this stuff!

Transferring to another HDD in the middle of the session isn't a good idea, as it'll take longer than the actual session since you're limited by the bandwidth of the harddisk - it would probably be better to go for somewhat fewer frames per section.

 

The old 2010 image was taken using a GSO 12" F/5 Newtonian, the new image with my newer 20" F/4 telescope, so it's not just the newer technologie, the nearly threefold increase in light collection area also helped ;)



#23 aeroman4907

aeroman4907

    Sputnik

  • -----
  • topic starter
  • Posts: 41
  • Joined: 23 Nov 2017

Posted 08 December 2017 - 10:13 AM

Good point Bart.  I think I might be able to get an external SSD drive that would work on USB 3 (3.1?)

 

20" f/4 - nice!



#24 Wildetelescope

Wildetelescope

    Viking 1

  • -----
  • Posts: 783
  • Joined: 12 Feb 2015
  • Loc: Maryland

Posted 08 December 2017 - 11:32 AM

Beautiful image JMD.  It will be interesting to see how processing video works.  There are a number of adjustments I can do in single shot RAW files in conversion to a 16 bit image file (nothing crazy, but to get some level of color to show up).  After images are stacked, I have a work around to convert the image file to a 64 bit file XIFS in PixInsight.  All my color saturation adjustments are performed in the 64 bit space.  Once I'm done with that, I save as a 16 bit TIFF and then make the color corrections in PS.  The PI purists would say do those color corrections in PI, but PS is so much more intuitive and powerful in this regard.  If I could do it in the 64 bit space, I would, but all that extra effort is not worth the return.  As it is, it takes some time in PS to get it just right.

 

Bart, those are amazing images!  What scopes were you using?  Regarding the computer, I have a very high end laptop with a good sized SSD for my business.  The SSD is not 1TB though, so I'll have to transfer files during about half way through a full moon capture to one of the mechanical drives and then finish imaging.  Now I've got to be disciplined today and get more work done on my real business that currently pays the bills.  I can get a little caught up in this stuff!

Thanks Aeroman!  I am pretty conversant in the aspects of image processing involving sharpening, deconvolution, etc..(I do microscopy in my day job), but Color adjustment is an aspect that I am really still learning about and have only dabbled with things like color channels, etc..(not much color in electron microscopy:-)   I agree that for this sort of thing, Photoshop is the gold standard.   I use Corel paintshop pro now, because I got angry with Adobe when they stopped maintaining their license servers for their legacy products, forcing everyone into their subscription model for purchasing things.     But the adobe product certainly is powerful, and handles things like 16 bit images, etc... better than the Corel product.   I have played around with Pixinsight's demo, but never pulled the trigger on getting a copy.   I just found that I could accomplish much of the same things with a suite of tools like Astra Image, Nebulosity, and Star tools, which I found a bit more intuitive to use.  To be sure, PI is very powerful, especially with its scripting capability, but for that level of effort, I might as well just go and teach myself programing in Matlab or Labview, two pieces of software I have access to through my job.  I also use the FIJI version of Image J which has a lot of powerful tools, although it is not necessarily as user friendly.   Lots of ways to skin the cat.:-)  I am continually amazed at what the average consumer can get their hands on for little or no money regarding imaging processing.  

 

Cheers!

 

JMD


  • aeroman4907 likes this

#25 Bart Declercq

Bart Declercq

    Viking 1

  • *****
  • Posts: 676
  • Joined: 21 Jan 2007
  • Loc: Haaltert, Belgium

Posted 09 December 2017 - 07:24 AM

Good point Bart.  I think I might be able to get an external SSD drive that would work on USB 3 (3.1?)

 

20" f/4 - nice!

That'll probably only work if your SSD is on a different USB-controller than the camera, otherwise you might be sharing bandwidth between the two which will kill high framerates (sorry for being a killjoy, but it's never nice to discover things like that *after* you've purchased stuff) - my observing laptop (a Lenovo T540p) had a second SATA port connected to the internal DVD-drive, I found an adapter kit which allowed me to replace the DVD with a second HDD, so I added a 1TB SSD - right now it's completely full, after only two observing sessions each less than 2 hours long. I used to keep my raw data backed up for possible reprocessing later, but I gave up as the volumes are just too large nowadays.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.







Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics