Monday 10 March 2025
In the vast expanse of radio astronomy, detecting transient phenomena has long been a challenge. The Square Kilometre Array (SKA), set to become the world’s largest radio telescope, will generate an unprecedented amount of data that requires efficient processing to identify these fleeting events. A team of researchers has developed a novel approach to tackle this issue by harnessing the power of graphics processing units (GPUs) and two intensity-sensitive image quality assessment methods: Low-Information Similarity Index (LISI) and augmented LISI (augLISI).
Radio astronomers typically use source detection algorithms to identify celestial transients in images. However, these algorithms can be inefficient when dealing with vast amounts of data from wide-field observations. The new approach uses GPUs to accelerate the processing of large images, achieving a significant speedup over traditional methods.
The researchers developed two image quality assessment methods: LISI and augLISI. These algorithms evaluate the similarity between input images by comparing pixel values. While both methods are intensity-sensitive, they differ in their focus. LISI is more sensitive to changes across all intensity levels, whereas augLISI focuses on significant variations.
The team tested their approach using simulated radio astronomical images generated with the Oxford’s Square Kilometre Array Radio-telescope simulator (OSKAR). They divided the images into tiles and applied the GPU-accelerated transient finders based on LISI and augLISI. The results showed that both methods effectively identified transients, but with different strengths.
LISI was highly sensitive to changes in all intensity levels, even subtle variations. This sensitivity is reflected in the wide range of LISI values across the image. On the other hand, augLISI focused on significant changes and reduced its sensitivity to minor variations, localising transients to tiles with fewer false positives.
The researchers demonstrated that their approach can be used for transient detection in radio astronomical images. The GPU acceleration enables efficient processing of large images, making it suitable for future applications such as the SKA. This novel approach has the potential to revolutionise the field of radio astronomy by enabling the rapid identification and characterisation of celestial transients.
The development of this technology is a significant step forward in the quest to understand the mysteries of the universe. As the SKA begins its operations, it will generate an enormous amount of data that requires efficient processing to unlock its secrets.
Cite this article: “Accelerating Transient Detection in Radio Astronomy with GPU-Powered Image Quality Assessment Methods”, The Science Archive, 2025.
Radio Astronomy, Square Kilometre Array, Gpu Acceleration, Transient Detection, Image Quality Assessment, Low-Information Similarity Index, Augmented Lisi, Oskar Simulator, Radio Astronomical Images, Celestial Transients
