The vast amount of data generated from ground- and space-based observations of the Sun in the last few decades has propelled the field of solar physics into the era of Big Data. Because of the sheer volume of this data, analysis solely by human observers is no longer possible. To ensure the quality of recorded images is good enough for further scientific analysis, scientists need an objective image-quality measure, especially for ground-based observations where clouds and other conditions can affect image quality. “As humans, we assess the quality of a real image by comparing it to an ideal reference image of the Sun,” explained Tatiana Podladchikova of Russia’s Skolkovo Institute of Science and Technology in a news item posted on the ‘Phys.org’ website. “For instance, an image with a cloud in front of the solar disk — a major deviation from our imaginary perfect image — would be tagged as a very low-quality image, while minor fluctuations are not that critical when it comes to quality. Conventional quality metrics struggle to provide a quality score independent of solar features and typically do not account for clouds,” said Prof. Podladchikova who used artificial intelligence (AI) to achieve human-like quality assessment together with three researchers from the University of Graz.
Using a neural network
Supported by the EU-funded SOLARNET project, the researchers developed a novel method for reliable image-quality assessment for ground-based full-disk solar observations. The method, which is described in a paper published in the journal ‘Astronomy & Astrophysics’, employs an unsupervised deep-learning approach that learns only from high-quality images. A neural network is used to learn the characteristics of high-quality observations and to detect deviations from them, consequently providing an objective image-quality score and reliably catching anomalies in the data. “In our study, we applied the method to observations from the Kanzelhöhe Observatory for Solar and Environmental Research and showed that it agrees with human observations in 98.5% of cases,” observed lead author Robert Jarolim of project partner University of Graz in the ‘Phys.org’. “From the application to unfiltered full observing days, we found that the neural network correctly identifies all strong quality degradations and allows us to select the best images, which results in a more reliable observation series. This is also important for future network telescopes, where observations from multiple sites need to be filtered and combined in real-time,” Jarolim went on to say. “Solar data delivery is the biggest project of our times in terms of total information produced. With the recent launches of groundbreaking solar missions, Parker Solar Probe and Solar Orbiter, we will be getting ever-increasing amounts of data offering new valuable insights. There are no beaten paths in our research. With so much new information coming in daily, we simply must invent novel efficient AI-aided data processing methods to deal with the biggest challenges facing humankind,” stated Prof. Podladchikova. The researchers’ novel method can provide reliable image-quality assessment in real time, without the use of any reference observations. According to the SOLARNET (Integrating High Resolution Solar Physics) study, the approach could also be applied to similar astrophysical observations and “requires only coarse manual labeling of a small data set.” For more information, please see: SOLARNET project website
SOLARNET, Sun, solar image, image quality, observation, artificial intelligence, data