Earth\u2019s atmosphere blurs images obtained even by the world\u2019s best ground-based telescopes due to shifting pockets of air. Although the blur seems harmless, it obscures the shapes of objects in astronomical images, which can lead to error-filled physical measurements that are essential for understanding the nature of our Universe. To overcome this, researchers at Northwestern University and Tsinghua University have adapted a well-known computer-vision algorithm used for sharpening photos, and applied it to astronomical images from ground-based telescopes<\/a> for the first time.<\/p>\n
The light emanating from distant stars, planets, and galaxies travels through Earth\u2019s atmosphere before it hits our eyes. The atmosphere blocks certain wavelengths of light and even distorts light before it can reach Earth. Clear night skies also affect light passing through it. That is why the best ground-based telescopes are located at high altitudes where the atmosphere is the thinnest.<\/p>\n
\u201cIt\u2019s a bit like looking up from the bottom of a swimming pool,\u201d Alexander said. \u201cThe water pushes light around and distorts it. The atmosphere is, of course, much less dense, but it\u2019s a similar concept.\u201d<\/p>\n
By studying the shape of galaxies, scientists can detect the gravitational effects of large-scale cosmological structures, which bend light on its way to our planet. This can cause an elliptical galaxy to appear rounder or more stretched than it really is. The atmospheric blur can warp the galaxy\u2019s shape. Thus, removing the blur allows scientists to collect accurate shape data.<\/p>\n
\u201cSlight differences in shape can tell us about gravity in the Universe,\u201d Alexander said. \u201cThese differences are already difficult to detect. If you look at an image from a ground-based telescope, a shape might be warped. It\u2019s hard to know if that\u2019s because of a gravitational effect or the atmosphere.\u201d<\/p>\n
Alexander and Tianao Li, an undergraduate in electrical engineering at Tsinghua University and a research intern in Alexander\u2019s lab, combined an optimisation computer-vision algorithm with a deep-learning network trained on astronomical images. Among the images, the team simulated data that matches the Rubin Observatory\u2019s expected imaging parameters. The resulting algorithm produced images with 38.6% less error compared to classic methods for removing blur and 7.4% less error compared to modern methods.<\/p>\n
When the Rubin Observatory opens next year, its telescopes will start a deep survey across a large portion of the sky. Because the researchers trained the computer-vision algorithm on data specifically designed to simulate Rubin\u2019s upcoming images, it will be able to help analyse the survey\u2019s highly anticipated data.<\/p>\n
For astronomers interested in using the tool, the user-friendly code and accompanying tutorials<\/a> are available online.<\/p>\n