New Images from Old Data

Joe DePasquale recently made astronomy news for his image of NGC 3621, a “pure-disk” galaxy hailed as a glorious example of a bulgeless spiral in a European Southern Observatory press release. His image ranked fifth in the ESO’s Hidden Treasures competition, in which the public was invited to take images from the ESO’s archives and process it into beautiful, poster-worthy showcases.

The 10 highest-ranked images in the Hidden Treasures Competition.
(Click here for a larger version of this slide show.)

Yet the most interesting story isn't the objects that DePasquale and others chose, but their experiences in learning to convert uncalibrated, gray-scale archival images into the impressive cosmic vistas they submitted. The ESO warned that taking on the Hidden Treasures challenge was not for the faint-hearted, a sentiment DePasquale echoes as he reminisces about the process.

DePasquale started with impressive credentials. He works at the Harvard-Smithsonian Center for Astrophysics as the science imager for the Chandra X-ray Observatory. Nevertheless the ESO’s archival data presented a new and daunting challenge. “I’d never dealt with raw optical data,” he says.

He first needed to learn the Unix-based program Image Reduction and Analysis Facility (IRAF), which he calls notoriously difficult to navigate. “I would maybe compare it to trying to assemble IKEA furniture without a manual,” he says.

Next step: choosing objects that would be worth the time and effort. “I had a browser open and was using Wikipedia to look through Messier objects. You can go through and say, ‘Okay, I think that looks like a good one.’” Then with luck, the ESO archive would have high-quality data on the chosen object.

Celestial Prominences
Joe DePasquale's image of NGC 3582, whose celestial prominences attracted Apple's interest and ranked 10th in the ESO's Hidden Treasures contest.
ESO/Joe DePasquale
But navigating the ESO archive was far from simple. “Every object was a new obstacle,” DePasquale says. The terabytes of ESO data are “kind of a mess,” making it difficult to find the appropriate calibration information. But the finished product made him feel it was worth it.

DePasquale is even getting a little recognition for an image of his that came in 10th in the contest: an Apple staffer contacted him to get his permission to use it in a new application the company is developing.

Want to learn more? Here's DePasquale’s slightly more technical rundown of what he did and how long it took:

“I would estimate that it took about 8 to 10 hours to process the first image. I'm estimating this from the time I downloaded the first image files, so it includes time spent searching for and downloading calibration files, setting everything up to reduce the data in IRAF, and then doing the data reduction.

"Once the data were reduced, I generated the calibrated images in each band and opened them up in what has become my power tool of astro-imaging: PixInsight. The individual images were registered, combined, color corrected, and stretched in PixInsight. I then took the combined image into Photoshop to tweak color intensities and clean up any residual artifacts (cosmic rays, bad pixels, etc). I probably shaved about 2 or 3 hours off of this time for the second image.”

For more information, see the ESO’s Hidden Treasures 2010 website and the corresponding Competition Flickr Group.