Automatic localization and characterization of mid-latitude ionospheric plasma structures from all-sky airglow images using deep learning framework

Chakrabarti, S. and Upadhyaya, J. and Rathi, R. and Yadav, V. and Patgiri, D. and Dixit, G. and Sunil Krishna, M.V. and Sarkhel, S. (2025) Automatic localization and characterization of mid-latitude ionospheric plasma structures from all-sky airglow images using deep learning framework. Advances in Space Research, 76 (4). pp. 2302-2314. ISSN 0273-1177

Text (Author_Final_Version)
Author_Final_Version.pdf - Accepted Version
Available under License Creative Commons Attribution.
Download (1MB)

Abstract

This paper brings forth a new automatic approach to determine the propagation parameters (horizontal velocity, propagation direction, and orientation) of mid-latitude ionospheric plasma structures (MIPS) using airglow images recorded by an all-sky airglow imager located at Hanle, Ladakh, India. The proposed approach is an amalgamation of two frameworks – a hybrid deep learning image segmentation model for localization along with automatic determination of parameters using the intensity minima of the MIPS. Designed in the form of a pipeline, the frameworks are executed sequentially. The propagation parameters obtained from the automatic method have been compared with the results of a previously implemented semi-automatic approach. Comparison between the two approaches revolves around the error involved, time complexity, and dependency on the morphology of the plasma structures. The results suggest that the proposed method can be adopted over the semi-automatic approach as it has less error, minimal dependency on the morphology of the structures, and less time-exhaustive.