Hubble’s true colours

With the advent of large observatories (on the ground and in space) and large detectors, astronomers have become accustomed to dealing with a vast amount of digital images of our universe. Be it stars, planets, galaxies or nebulae, today's most stunning views of the heavens are stored in digital images of outstanding quality, unprecedented sharpness and ever growing size.

Scientists need practical and efficient tools to deal quickly and reliably with these large datasets and IDL is the language of choice for many of them. Guido De Marchi, an astronomer with the European Space Agency, has been using IDL regularly in his work since the very first days in Hubble's history, in 1990, when a manufacturing flaw discovered in the telescope's primary mirror made all images look fuzzy and blurred. "The ease of use, versatility and speed of IDL made it the right tool to work with Hubble's first images that needed a lot of careful processing," he explains. A young student in those days, De Marchi worked with his colleagues on methods to extract scientific information from the blurred images without compromising the integrity of the data, by means of a technique called deconvolution. "The prospect of having to do all the coding in Fortran or C was terrifying, but then I discovered IDL. It was love at first sight."

In 1993 astronauts refurbished Hubble and installed new corrective optics. The telescope has since been delivering pictures of the highest resolution and deconvolution is no longer necessary. But for De Marchi this did not mean abandoning IDL. In his research on stellar clusters, he makes routinely use of Hubble's data to study the distribution and mass of their stars and IDL remains his preferred tool for measuring accurate positions, brightness and colours of thousands of objects in the images. He uses this information to understand how stars form and how systems comprising millions of them evolve with time.

But IDL lends itself equally well to combining images taken through different filters and create true-colour views of the universe. Astronomers use filters of different colours to probe the properties of stars and galaxies at different wavelengths. For instance, blue and ultraviolet filters are particularly suited to studying very young and massive stars, while red filters are most sensitive to old stars and to small stars still embedded in their natal clouds. "We usually have observations of the same region of sky in three or more wavelength bands and IDL makes it easy to combine them together into a single true-colour image," says De Marchi.

The picture here below shows a region of violent star formation in the Large Magellanic Cloud, the nearest neighbour galaxy to our Milky Way and one of the jewels in the Southern sky, located at a distance of about 165,000 light years. Here three Hubble exposures have been combined. They were taken respectively through an ultraviolet, a blue and a red filter. IDL allows for the seamless rescaling of the three image components, for instance to account for different exposure times or brightness of the objects, and their combination into a single array with the proper colour table for each RGB channel.

(image credit NASA, ESA and Guido De Marchi)

Unfortunately, not everything is perfect. The limited dynamic range of the Hubble camera used in this investigation implies that bright stars tend to saturate in long exposures aimed at detecting the faintest objects. This explains why the most luminous stars in this picture look like doughnuts, with a hole in the centre. De Marchi is not worried, though, since bright stars are easy to study and measure on shorter exposures or by using methods such as deconvolution.

True-colour pictures give scientists better insight into the physics of the processes that they are studying. For instance, the distribution of red and blue stars in the image is not the same. While faint red stars are uniformly spread over the entire field, bright white-blue stars are clumped together, particularly near the lower left corner of the frame. "Bright blue stars are young and we see them where they just formed," explains De Marchi, "but faint red objects can be one or two billion years old and have likely wandered a long way from their birthplace."

With the Hubble archives open for anybody to use, scientists or computer pundits alike, and the availability of simple routines to read astronomical images in their native FITS format, IDL users can unleash their creativity and easily develop their own celestial gallery. After all, besides their immense scientific value, Hubble's pictures are a true candy to the eye.