Real-ESRGAN and SWIN-U-NET Applications in Enhancing NASA Imagery for VR Solar System Models

Immersive Virtual Reality (VR) models of the solar system hold immense potential for education and scientific exploration. However, limited resolution and artifacts in NASA imagery often hinder the visual fidelity and scientific accuracy of these VR experiences. This paper explores the application of two cutting-edge image super-resolution algorithms, Real-ESRGAN and SWIN-U-NET, to significantly enhance the spatial resolution of NASA imagery for VR solar system models. A comparative analysis of their performance on several NASA datasets is presented in this research., encompassing planetary surfaces, celestial bodies, and astronomical phenomena. The analysis focuses on metrics relevant to VR experiences, including resolution gain, noise reduction, artifact removal, and preservation of scientific detail. This proposed technology produces remarkably high-resolution textures, perfect for adorning 3D objects in the proposed future Galaxy VR model. These enhanced images lie at the heart of an immersive and captivating VR experience, inviting users to journey through the solar system with unprecedented visual clarity and detail. By crafting hyper-realistic, captivating VR simulations of celestial objects and phenomena, this technology holds the potential to revolutionize astronomy education and exploration, especially within the solar system and its diverse range of phenomena. In this paper, Peak signal-to-noise ratio (PSNR) and the structural similarity index measure (SSIM) methods were used for calculating the accuracy of NASA enhanced imagery. The Phenomena datasets showed the highest improvement, with average accuracy rising from 34% to 91%. Planetary surfaces also saw a significant boost, reaching 39% and 90%. While celestial bodies showed lower improvement, reaching 33% and 83% respectively, the results still represent a valuable step forward. Finally, The discussion delves into the implications of these findings for the development of next-generation VR solar system models, emphasizing the potential for heightened realism, improved scientific accuracy, and ultimately, a more profound understanding of our cosmic neighborhood.