Visualizing Hubble Data

Zolt Levay, levay@stsci.edu

The magical combination of art and science in Hubble color images has engaged a world-wide audience in astronomical exploration and discovery. This phenomenon has amazed and deeply gratified everyone working with Hubble, but perhaps most keenly the News Team in the Institute’s Office of Public Outreach, which produces the pictures. This article is a look behind the scenes at how we transform Hubble science data into compelling color images.

Of course, the primary purpose of images taken by Hubble is to do science. Images are obtained with one of many possible combinations of filters available in Hubble’s cameras. The intensity of the light source is faithfully recorded in these images and can be cleverly utilized in a variety of ways to help answer the scientific questions that led to them being obtained in the first place. The Institute’s news team distributes press releases on scientific advances, which frequently include pictures, through news media, educators, and the Internet.

A second important quality of the some of the images from Hubble is their undeniable artistic appeal, especially when multiple filters have been used, and a color image can be produced.  These high-profile images act as a magnet for the underlying science. Each new color image demands careful treatment due to both the exquisite quality of the data and the imperative to be faithful to nature. We like to think that the resulting color pictures are as good as they can be, limited only by the sky and the telescope.

In the following sections, we lay out the essentials of translating Hubble data into color pictures. Some of the steps are as old as photography. Others—including the first step, capturing the data with Hubble’s state-of-the-art cameras—are at the cutting edge of technology for image processing. We hope these explanations lead to deeper appreciation of the sheer beauty of Hubble pictures.

 

Monochrome to color

Hubble’s science instruments do not produce color images directly, of course. In each exposure, a filter restricts the spectrum of detected light. The astronomer selects the filter to highlight (or suppress) particular objects or physical processes occurring in the field of view. Often the science is in a difference in brightness between exposures with different filters. Because we know the spectral transmittance of each filter exactly, we can extract and reconstruct color information from a series of monochrome images of the same field taken through a variety of filters.

Even though the filters selected by the astronomer are rarely red, green, and blue—the primary additive colors—we generally use those colors for reconstruction. We assign red to the lowest energy image, blue to the highest, and green to the intermediate energy. While any set of hues might be used, the choice of the primary colors maximizes the color gamut of the reconstructed picture. We sometimes vary this approach to improve aesthetics, shift colors for practical reasons, or enhance particular features. In any case, the chosen blend always maintains a relationship to true physical phenomena in the scene. We call such schemes “representative” color rather than false- or pseudo-color.

 

Intensity scaling

After routine data reduction, scaling the intensities is the first step in preparing monochrome images for combining. A flexible tool for this is FITS Liberator, which is free, special-purpose software developed at the European Space Agency’s Space Telescope European Coordinating Facility. This tool provides flexibility in applying intensity-transform functions, including clipping. Using it, we scale each monochrome image to optimize the tonal range. (All subsequent steps for Hubble color pictures are accomplished using industry-standard, image-editing software.)