Tuesday, December 07, 2010

Usefully consuming time

kw: science, astronomy, public service

Want to do some real astronomy, something helpful? You have something no computer program can duplicate, a finely tuned recognition ability. Assuming you aren't blind (and thus using some kind of assistance program to "read" this text), your vision system can quickly discriminate objects from background clutter and distinguish their shapes.

One harvest of the various telescopic surveys, including the Sloan Digital Sky Survey (SDSS) and the Hubble Space Telescope (HST) program, is a huge archive of deep space images of millions of objects. Learning how the Universe works requires classifying large numbers of these objects. While a computer program can scan a digital image and locate objects, and even discern a star from a galaxy with quite good accuracy, it usually cannot determine whether a particular galaxy is elliptical, spiral, irregular or something else. Guess what? With a minimum of preparation, you can!

Three years ago a team of astronomers launched a project called Galaxy Zoo, a web-based classification program for the galaxies identified in SDSS images. They hoped to have a million galaxies classified by several persons each over a span of a few years. They got a huge public response, and received tens of millions of classifications, an average of fifty "votes" each for a million galaxies, in about five months.

As the astronomers learned more about galaxies in the early universe, they learned better questions to ask. The project has been expanded to the deeper images found in the HST archives, and is currently the Galaxy Zoo Hubble, one of several Zooniverse projects. I encourage everyone who can, bookmark the link, register, and learn to classify galaxies. The one-web-page tutorial is a great introduction to the subject. What follows introduces some of the concepts, because I've been having great fun, having classified my first 400 galaxies.

Click on this image to see it full size. It shows the control panel for classifying HST galaxies. You answer the Zoo's questions by clicking on the buttons on the right. As this screen shot shows, the first question is to determine, is this a smooth (that is, elliptical) galaxy, or something else? In the "recent" Universe, that within about a billion light years, most galaxies are elliptical. However, the public perception of galaxies is shaped more by the beautiful spiral galaxies.

A word of caution. There are few "pretty pictures" in the GZH archive. The focus of cosmologists' interest is the early Universe, so the galaxies of interest are very far away, and therefore appear rather small in HST images.

This is one of the better images of a spiral galaxy. Two spiral arms are visible, and there is a small central bulge, the orange spot near the center. Questions asked by GZH lead you to describe each feature. For this one, I selected "Features or Disk", "Not clumpy", "Not an edge-on spiral" (Yes to this one leads one through different questions), "No Bar", "Spiral", "Medium tightness", "2 Arms", "Just noticeable bulge".

There is also a question, "Anything odd?", which is where you'd designate whether the galaxy is irregular, or a merger of two or more, or has a prominent dust lane (usually seen only in spirals that are nearly edge-on), and a few other features.

During classification, you can also see the image as a negative, with dark features on a white background. This is often useful for counting the spiral arms when a galaxy has a more ambiguous appearance.

Here is one of the better images of an elliptical galaxy. This one is part of a cluster (you can see one of equal size to the upper left), and this cluster is in front of an enormous cluster of more distant galaxies. However, you are asked to classify only the object at the center of the image.

Ellipticals are simple; for this image I answered "Smooth", "Medium oval", and "Nothing odd". A computer program would need to be very smart to figure out that the five colored dots "inside" the large elliptical galaxy's image are actually galaxies situated much farther away, seen through it.

With one exception, everything you see in this image is a galaxy. That exception is the overexposed image of a star at the lower left. If GZH presents such an image to you, you'd click "Star or Artifact". I've also seen various artifacts such as a "dotted line" caused by something in motion crossing the field of view during the exposure: HST discovered lots of asteroids that way.

This elliptical galaxy happens to have a distinct central condensation. Many do, but not all. One way to distinguish an elliptical galaxy with a central condensation from a face-on disk galaxy is that the disk and central bulge are different colors, while all of an elliptical galaxy will be the same color.

Many images of the farthest objects show galaxies that are still forming from smaller objects. They have a clumpy appearance like this. For this image I answered "Features...", "Clumpy", "More than 4 clumps", "Clustered", "No distinctly brightest clump", "Not symmetrical", "Not embedded in something larger", and "Nothing odd".

The proctors of the GZH web site are clever. One of the answers to how the clumps are arranged is "Spiral pattern". Early spiral galaxies have bright star-forming regions strung along the spiral arms, and if we cannot see the whole galaxy, all we see are those bright spots. But they outline the spiral shape, so such an image winds up being provisionally classified as a spiral galaxy.

It is remarkable how consistent the classification of most objects is. Each one is presented to a number of people, so the professional astronomers only have to take a second look at the ones that get a scattering of widely divergent classifications. This last image might be such a one.

If you are feeling helpful and have a little time on your hands, give the GZH project a try.


shelyad said...

Excellent Blog-Michael Druck-shelyad@gmail.com

shelyad said...