Search papers, labs, and topics across Lattice.
This survey paper reviews deep learning-based methods for detecting and segmenting solar corona structures, specifically coronal holes and active regions, due to their importance in space weather forecasting. It highlights the superior performance of modern deep learning architectures compared to traditional methods in this domain. The paper concludes by discussing potential applications in space weather operations and identifying open research challenges.
Deep learning is making inroads in space weather forecasting, but this survey reveals the specific architectures and open challenges in applying them to solar corona structure detection.
Image segmentation and detection of solar corona structures are a critical part of space weather applications. In recent years, deep learning-based techniques have proven to be effective for various automatic solar image processing tasks. Modern methods based on novel deep learning architectures often perform better in comparison to other traditional machine learning or computer vision methods in various application domains. This article provides a review of the current approaches and applications of deep learning methods in the specific area of solar image data analysis. In particular, we focused on detection and segmentation techniques of coronal holes and active regions due to their dominant role in solar-terrestrial interactions. The article introduces object detection and segmentation techniques and subsequently provides an overview of their application in the detection and segmentation of solar corona structures. Finally, we discuss the potential applications of the methods in space weather operation services, emphasizing their adaptability to handle different solar data, and identify some of the open research issues in the analyzed area.
