Jump to content


Photo

Why aren't there color POSS images?

  • Please log in to reply
7 replies to this topic

#1 derangedhermit

derangedhermit

    Apollo

  • -----
  • Posts: 1150
  • Joined: 07 Oct 2009
  • Loc: USA

Posted 10 July 2013 - 03:19 AM

For example POSS I has red and blue plates, the southern sky survey has red and blue, and POSS II has red, blue, and near IR. They are all available separately, and I have found one amateur taking a very few of them and combining them to make very pretty pictures; but in general I don't see them combined to form color images. They are mostly all available in digital form now. Is it assumed there is no scientific value in looking at them together?

Lee

#2 Tony Flanders

Tony Flanders

    Voyager 1

  • *****
  • Posts: 11179
  • Joined: 18 May 2006
  • Loc: Cambridge, MA, USA

Posted 10 July 2013 - 05:32 AM

For example POSS I has red and blue plates, the southern sky survey has red and blue, and POSS II has red, blue, and near IR. They are all available separately, and I have found one amateur taking a very few of them and combining them to make very pretty pictures; but in general I don't see them combined to form color images. They are mostly all available in digital form now. Is it assumed there is no scientific value in looking at them together?


You can't form true color images from red and blue alone. You have to guess about green (usually formed by averaging red and blue), which is the dominant color to the human eye.

Nonetheless, many people combine POSS plates to form colored images. Sean Walker and I do it all the time for Sky & Telescope magazine, and Wikisky does it automatically, on the fly.

There are lots of gotchas. Often a particular plate is missing or defective, and the exposures are fairly variable. Combining POSS plates well is a bit of a fine art.

Tony Flanders
Associate Editor, Sky & Telescope

#3 opticsguy

opticsguy

    Messenger

  • -----
  • Posts: 416
  • Joined: 02 Feb 2009
  • Loc: Washington State

Posted 10 July 2013 - 05:53 AM

You can also find an interesting discussion of this subject in the book "Planetfall" by Michael Benson. A fabulous book!!

#4 bicparker

bicparker

    Vanguard

  • *****
  • Posts: 2363
  • Joined: 07 Feb 2005
  • Loc: Texas Hill Country

Posted 10 July 2013 - 11:34 AM

Two reasons:
1. Before 1960, as everyone knows, all of the world was black and white, so even though they took "red" and "blue" images, they all came out black and white. Attempts to colorize the DSS have been attempted by Ted Turner, but this has met with extreme disdain from the astronomical community.
2. What Tony F. said.

I've done a few plate combinations synthesizing the green and I know of several who have done this. It is not that hard and you can get some interesting results, though, like Tony mentioned, some of the plates have some problems. One problem I ran into a lot had to do with areas with bright objects or stars or stars near objects. One case in particular, in Fornax, the star right by NGC 1404 is strong in the red making it very bloated on the plate and going into the galaxy, while the blue shows a much smaller star. Synthesized green results can look odd. It just doesn't combine well and requires a lot of post work to make it look reasonable.

These plates were not really meant to be combined, but rather they were filtered and calibrated to yield survey data on a stand alone basis for that filter run. That doesn't mean that you can't derive data from the combined images, but the yield would probably be more productive testing against a narrow hypothesis and using other data, as well (such as X-Ray or mid to far IR, for instance).

#5 derangedhermit

derangedhermit

    Apollo

  • -----
  • Posts: 1150
  • Joined: 07 Oct 2009
  • Loc: USA

Posted 12 July 2013 - 03:41 AM

Two reasons:
1. Before 1960, as everyone knows, all of the world was black and white, so even though they took "red" and "blue" images, they all came out black and white. Attempts to colorize the DSS have been attempted by Ted Turner, but this has met with extreme disdain from the astronomical community.
2. What Tony F. said.

I've done a few plate combinations synthesizing the green and I know of several who have done this. It is not that hard and you can get some interesting results, though, like Tony mentioned, some of the plates have some problems. One problem I ran into a lot had to do with areas with bright objects or stars or stars near objects. One case in particular, in Fornax, the star right by NGC 1404 is strong in the red making it very bloated on the plate and going into the galaxy, while the blue shows a much smaller star. Synthesized green results can look odd. It just doesn't combine well and requires a lot of post work to make it look reasonable.

These plates were not really meant to be combined, but rather they were filtered and calibrated to yield survey data on a stand alone basis for that filter run. That doesn't mean that you can't derive data from the combined images, but the yield would probably be more productive testing against a narrow hypothesis and using other data, as well (such as X-Ray or mid to far IR, for instance).


1. I have a divided opinion; some movies are just fine in B&W, thank you for leaving it alone, and little is added by color. Others, you watch in B&W and think "a color version would have been nice"; and for these latter, I think mostly a decent enough job has been done.

2. The last of what you wrote is what I suspected. There may be a few odd cases where some astronomer might want the R&B superposed for some specific reason, but not often enough to warrant doing them all.

We are often surprised by the universe, though, and I would wager that if you sat a few folks down for a month to look at combined images, you would get a number of "well, isn't this one interesting..."

Thanks for the answers, folks.
Lee

#6 Tony Flanders

Tony Flanders

    Voyager 1

  • *****
  • Posts: 11179
  • Joined: 18 May 2006
  • Loc: Cambridge, MA, USA

Posted 12 July 2013 - 06:15 AM

There may be a few odd cases where some astronomer might want the R&B superposed for some specific reason, but not often enough to warrant doing them all.


Actually, color images have great scientific value. For instance, in a typical star field, you can pick out red giants and OB dwarfs at a glance. You can tell spiral galaxies from ellipticals by their characteristic colors, and estimate the intervening dust by the reddening.

Most useful of all are false-color images that enhance the all-important hydrogen-alpha line.

There's a reason why the Sloan Digital Sky Survey, arguably the successor of POSS, presents color images as one of its primary interfaces.

However, the SDSS was designed to record color, with multiple cameras using different filters operating simultaneously. The POSS plates were taken at different times in different wavelengths, so you inevitably end up with blue or red satellites and meteors even in the best cases. Variable stars, which are both numerous and important, may have been recorded at maximum in one wavelength and minimum in another.

Despite those reservations, however, the Wikisky colorization of POSS-II is an invaluable resource.

#7 bicparker

bicparker

    Vanguard

  • *****
  • Posts: 2363
  • Joined: 07 Feb 2005
  • Loc: Texas Hill Country

Posted 12 July 2013 - 08:27 AM

I know this is a bit of a tangent from the OP's question, but I wanted to say this in any case.

Probably one of the neatest features of the POSS is something we are just beginning to see. As it gets older, it's value as set of historical data points of stuff in space increases. So much of astronomy is not just analyzing the state of things in space, but how they change over time. And with the longer breadths of time, the better we can draw conclusions or even make new hypotheses with those actions that take longer time frames.

The fact that we can make some synthesized color images, albeit with some limitations brought on by the timing and other issues that Tony mentioned, adds to this historical value. Those color comparisons add a new dimension to the comparative value that we don't have with many older data elements.

It is an exciting use of resources like the POSS. We have already seen some very significant uses of the POSS in this manner and I expect that will only increase over time.

#8 GlennLeDrew

GlennLeDrew

    Voyager 1

  • *****
  • Posts: 10647
  • Joined: 17 Jun 2008
  • Loc: Ottawa, Ontario, Canada

Posted 12 July 2013 - 03:50 PM

A group in Japan (Dobashi, et al) fairly recently scanned the POSS (and its southern counterpart) to compile a new and more uniform all-sky catalog of optical dark nebulae and cloud cores. (The detected features are correlated with previously catalogued objects.) The huge FITS file which maps visual extinction--and which is freely available--merits hours of study.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics