Researchers at Northwestern University in the United States and Tsinghua University in Beijing have built a neural network that smartly removes blur from space images. The study, which showcases advances in image restoration, appears in scholarly records and is also shared via arXiv preprint archives.
Ground-based telescopes often yield blurry photographs. The culprit is the atmosphere, where moving air pockets distort light as it travels from distant stars, planets, and galaxies to Earth. The result is softer details and less sharp edges than what space reveals.
To tackle this challenge, engineers combined a refined artificial intelligence algorithm with a deep learning system. The goal was to accelerate the sharpening process while delivering more realistic and striking results. In tests, the new model achieved noticeably clearer images and reduced distortions compared with conventional deblurring approaches. The open-source code and accompanying documentation available in the project repository invite researchers worldwide to experiment and build on the method.
Alongside these developments, former NASA scientists presented imagery of a colossal solar event described as a 14-Earth-mass phenomenon. The so-called solar hurricane appeared near the Sun’s north pole around March 15 and evolved rapidly until it dissipated on March 18. The event involved a surge of plasma released into space, a reminder of the dynamic activity on and around the Sun. Astronomers noted that the ejected plasma would not reach Earth, but it could interact with the Moon and surrounding space environment.