Solar Images Research Articles

Column fixed-pattern noise removal in solar images using two-way filtering

Column fixed-pattern noise removal in solar images using two-way filtering

Total Solar Eclipse White-light Images as a Benchmark for Potential Field Source Surface Coronal Magnetic Field Models: An In-depth Analysis over a Solar Cycle

Potential field source surface (PFSS) models are widely used to simulate coronal magnetic fields. PFSS models use the observed photospheric magnetic field as the inner boundary condition and assume a perfectly radial field beyond a “source surface” (R ss). At present, total solar eclipse (TSE) white-light images are the only data that delineate the coronal magnetic field from the photosphere out to several solar radii (R ⊙). We utilize a complete solar cycle span of these images between 2008 and 2020 as a benchmark to assess the reliability of PFSS models. For a quantitative assessment, we apply the Rolling Hough Transform to the eclipse data and corresponding PFFS models to measure the difference, Δθ, between the data and model magnetic field lines throughout the corona. We find that the average Δθ, 〈Δθ〉, can be minimized for a given choice of R ss depending on the phase within a solar cycle. In particular, R ss ≈ 1.3 R ⊙ is found to be optimal for solar maximum, while R ss ≈ 3 R ⊙ yields a better match at solar minimum. Regardless, large (〈Δθ〉 > 10°) discrepancies between TSE data and PFSS-generated coronal field lines remain regardless of the choice of source surface. However, implementation of solar-cycle-dependent R ss optimal values does yield more reliable PFSS-generated coronal field lines for use in models and for tracing in situ measurements back to their sources at the Sun.

Automated High-Precision Recognition of Solar Filaments Based on an Improved U2-Net

Solar filaments are a significant solar activity phenomenon, typically observed in full-disk solar observations in the H-alpha band. They are closely associated with the magnetic fields of solar active regions, solar flare eruptions, and coronal mass ejections. With the increasing volume of observational data, the automated high-precision recognition of solar filaments using deep learning is crucial. In this study, we processed full-disk H-alpha solar images captured by the Chinese H-alpha Solar Explorer in 2023 to generate labels for solar filaments. The preprocessing steps included limb-darkening removal, grayscale transformation, K-means clustering, particle erosion, multiple closing operations, and hole filling. The dataset containing solar filament labels is constructed for deep learning. We developed the Attention U2-Net neural network for deep learning on the solar dataset by introducing an attention mechanism into U2-Net. In the results, Attention U2-Net achieved an average Accuracy of 0.9987, an average Precision of 0.8221, an average Recall of 0.8469, an average IoU of 0.7139, and an average F1-score of 0.8323 on the solar filament test set, showing significant improvements compared to other U-net variants.

Removing cloud shadows from ground-based solar imagery

The study and prediction of space weather entails the analysis of solar images showing structures of the Sun’s atmosphere. When imaged from the Earth’s ground, images may be polluted by terrestrial clouds which hinder the detection of solar structures. We propose a new method to remove cloud shadows, based on a U-Net architecture, and compare classical supervision with conditional GAN. We evaluate our method on two different imaging modalities, using both real images and a new dataset of synthetic clouds. Quantitative assessments are obtained through image quality indices (RMSE, PSNR, SSIM, and FID). We demonstrate improved results with regards to the traditional cloud removal technique and a sparse coding baseline, on different cloud types and textures.

The Solar Aspect System of the Hard X-ray Imager Onboard ASO-S

The ASO-S/HXI is a bi-grids modulating instrument for solar hard X-ray imaging, whose collimator contains 91 pairs of tungsten grids. Since the solar disk is invisible in hard X-rays, a Solar Aspect System (SAS) is required to provide the pointing of hard X-ray imager (HXI) for locating X-ray sources on the solar disk. In addition, the knowledge of the alignment and relative twist of the corresponding front–rear grid pairs is important for image reconstruction as well as locating flares. Therefore, the SAS system was designed to monitor the alignment status of HXI grids and to provide the pointing direction of the HXI collimator with two subsystems DM and SA during the whole life cycle of HXI. DM measures the centroids of the front frosted glasses and the solar disk. SA images the Sun and provides precise relative locations of the solar disk center. Both work in the visible light of 565–585 nm. With all the data together, we can solve with an inversion algorithm the alignment status of the front and rear grids, the relative twist, and the pointing direction. We tested and validated the SAS design with both the simulation model and ground coordinate measuring machine. Here we present the detailed system design, the testing results, the inversion algorithm, and the in-orbit status of the SAS. Currently, the SAS has realized the rotational measurement accuracy of about 4 arcsec, and a translational measurement accuracy of about 15 μm, and the SAS pointing data has been used in both imaging calibration for flare locations and imaging corrections for the platform drifting effect. The high-cadence precise measurement (better than 0.3 arcsec) of the pointing will allow the study of source locations at different energies and therefore help us to understand electron acceleration and transportation in flares.

Parallel Analogy Method for Measuring FRD of the FASOT-IFU

A fiber integral field unit (IFU) was chosen for the Fiber Array Solar Optical Telescope (FASOT) to obtain high-spectral-resolution polarization solar spectral images. This IFU features a dual-array-end design, complemented by a dozen slit ends. Each consists of 2 sub-slits corresponding to the dual array ends. In order to accurately and quickly measure the focal ratio degradation of 8064 fibers in FASOT-IFU, a parallel analogy method (PAM) for FASOT-IFU was proposed. This method primarily relies on characterizing the focal ratio based on the full-width at half-maximum (FWHM) of the fiber’s output spot, and the linear relationship between the output focal ratio of the spot and the FWHM. By precisely calculating the focal ratio of the positioning fiber in the FASOT-IFU and treating it as a reference, the output focal ratio of the working fiber can be obtained by comparing the FWHM of the working fiber spot and the positioning fiber spot at the same emission distance. The average value of the output focal ratio by PAM corresponding to the 4032 fibers from Array A of the IFU is 6.0 and the standard deviation is 0.6. The values corresponding to the Array B are 6.2 and 0.40, respectively. The mean value exceeds the contractually specified F/5.5, and the proportion of optical fibers with a focal ratio greater than 5.5 accounts for 88% of the total number of optical fibers, meeting the design requirements of the FASOT system.

Study of changes in the pulsation period of six Cepheids variable

We study the period change of six Cepheids using 19376 accurate flux observations of the Solar Mass Ejection Imager (SMEI) onboard the Coriolis spacecraft. All observations for the six Cepheids have been derived as templates for each star, independent of the specific sites utilized to establish and update the O–C values. Sometimes, sinusoidal patterns are superimposed on the star’s O–C changes, which cannot be regarded as random fluctuations in the pulsation period. Random period changes were detected and computed using Eddington’s and Plakidis’s approaches. A comparison of the observed and predicted period change reveals a good agreement with some published models and a very substantial divergence with others. Between the reported period change and that estimated by the current technique, a linear fit with a correlation coefficient of 90.08 percent was obtained. The temporal rate of period change in Cepheid stars might be connected to how well these stars’ mass losses are known today.

The Large Imaging Spectrometer for Solar Accelerated Nuclei (LISSAN): A next-generation solar gamma-ray spectroscopic imaging instrument concept

We present the Large Imaging Spectrometer for Solar Accelerated Nuclei (LISSAN), a new solar dedicated satellite instrument concept. LISSAN relies on an indirect Fourier imaging technique valid over an energy range from 40 keV up to 100 MeV. Spatial information is encoded into 15 moiré patterns by 15 pairs of slightly offset grids (bigrids) separated by a fixed distance enabling a predicted spatial resolution of 10”. The time, location, and energy of each incoming photon is recorded via a pixelated gadolinium aluminium gallium garnet (GAGG) crystal scintillator detector placed in alignment with each bigrid, therefore providing simultaneous imaging and spectroscopy from the same imaging system. X-ray and gamma-ray emission are key diagnostics of electron and ion acceleration, respectively. However, despite being a fundamental process that occurs throughout the Universe, particularly in the solar atmosphere, only one resolved gamma-ray image of ion acceleration in a solar flare has ever been achieved. LISSAN will shed new light on this process by providing spectral resolution better than 1.5% FWHM at 6.1 MeV, an imaging effective area at 2.2 MeV of 100 cm2 (more than 25 times greater than past missions such as RHESSI) and a 10 s cadence. Thanks to these significant advances over the previous satellite-based solar detectors, LISSAN will provide reliable imaging and the spectral characterization of both electron and ion acceleration in solar eruptive events simultaneously for the first time, enabling to it to answer several important open questions regarding solar particle acceleration and the initiation of space weather events.

Triangulation of Hard X-Ray Sources in an X-Class Solar Flare with ASO-S/HXI and Solar Orbiter/STIX

HXI on ASO-S and STIX onboard Solar Orbiter are the first simultaneously operating solar hard X-ray imaging spectrometers. ASO-S’s low Earth orbit and Solar Orbiter’s periodic displacement from the Sun–Earth line enables multi-viewpoint solar hard X-ray spectroscopic imaging analysis for the first time. Here, we demonstrate the potential of this new capability by reporting the first results of 3D triangulation of hard X-ray sources in the SOL2023-12-31T21:55 X5 flare. HXI and STIX observed the flare near the east limb with an observer separation angle of 18°. We triangulated the brightest regions within each source, which enabled us to characterise the large-scale hard X-ray geometry of the flare. The footpoints were found to be in the chromosphere within uncertainty, as expected, while the thermal looptop source was centred at an altitude of 15.1 ± 1 Mm. Given the footpoint separation, this implies a more elongated magnetic-loop structure than predicted by a semi-circular model. These results show the strong diagnostic power of joint HXI and STIX observations for understanding the 3D geometry of solar flares. We conclude by discussing the next steps required to fully exploit their potential.

Performance of a Prototype Heliostat Having a Twisting Mechanism to Maintain Focus Throughout the Day

A prototype heliostat has been designed and built with a rectangular reflector whose shape is altered automatically through the day, to maximize concentration at the receiver. The shape changes needed to form solar disc images, which give the highest possible concentration, can be realized by twisting the reflector from its corners, provided that a target-oriented, dual-axis mount is used. Then a cam mechanism connected to the second (cross) axis drive, turning with the angle of incident sunlight, may be used to twist the reflector as needed. Our prototype reflector comprises a single glass mirror, 2.4 m x 3.3 m, attached to a steel frame. Four diagonal back struts extend from the central mechanism out to the rectangular frame corners. The glass mirror of the prototype is attached to the steel frame by 58 screw actuators. Before twisting the frame, the glass shape is adjusted to the biconic shape needed to form a 1 m diameter disc image on a 113 m distant target when sunlight is incident at 60°. To form disc images over the full range of angles of incidence from 0° (normal incidence) to 70°, the struts push the corners up and down by up to 17 mm. A reflectometry metrology system has been used to set the 58 adjustment screws for the initial shape to an accuracy of ≤ 0.6 mrad, and to measure the accuracy of the different twisted shapes. The prototype will be tested at NSTTF early next year.

Informing solar blind radioluminescence imaging through a calibrated spectrum

While direct radiation detection methods offer great insight into the origin of ionizing particles and photons, their use to locate contaminated areas or concealed radioactive sources can lead to undue exposure of personnel and equipment to ionizing radiation or the potential for contamination. These same sources induce ultraviolet (UV) optical photon fluorescence in air – a process referred to as radioluminescence – that may be imaged from low dose regions over larger attenuation lengths than ionizing radiation. However, most optical detection methods are limited to low lighting conditions to image the more abundant ultraviolet-A (UV-A) photons. To extend this capability to room light or daytime conditions, the solar blind region (Ultraviolet-C (UV–C), <280 nm) can be tapped. Though the emission yield of UV-C photons is roughly two orders of magnitude lower than that in the UV-A regime, the UV-C offers dramatic improvements in signal-to-noise ratios under bright lighting conditions due to decreased background interferences. The yield of specific UV-C lines, if present in the literature at all, varies widely, which has a large impact in modeling and analyzing standoff UV-C measurement scenarios. Thus, we have captured improved radioluminescence spectra over 250–400 nm and identified observed emission peaks in ambient air. Many of the UV-C photons produced by ionizing radiation excitation result from high-energy, molecular nitrogen Gaydon-Herman transitions which have had limited study to date for this application. Relating these findings to published UV-A yields, we estimate emissions between 0.15 and 0.19 photons/MeV over 250–280 nm. We also use a commercial corona-discharge imaging camera to demonstrate outdoor UV-C radiation mapping of alpha and gamma emitters from 50 to 75 m standoffs, respectively. The imaged “counts” are compared to optically modeled values and show the same trend over distance.

Predicting sunspot number from topological features in spectral images I: Machine learning approach

This study presents an advanced machine learning approach to predict the number of sunspots using a comprehensive dataset derived from solar images provided by the Solar and Heliospheric Observatory (SOHO). The dataset encompasses various spectral bands, capturing the complex dynamics of solar activity and facilitating interdisciplinary analyses with other solar phenomena. We employed five machine learning models: Random Forest Regressor, Gradient Boosting Regressor, Extra Trees Regressor, Ada Boost Regressor, and Hist Gradient Boosting Regressor, to predict sunspot numbers. These models utilized four key heliospheric variables — Proton Density, Temperature, Bulk Flow Speed and Interplanetary Magnetic Field (IMF) — alongside 14 newly introduced topological variables. These topological features were extracted from solar images using different filters, including HMIIGR, HMIMAG, EIT171, EIT195, EIT284, and EIT304. In total, 60 models were constructed, both incorporating and excluding the topological variables. Our analysis reveals that models incorporating the topological variables achieved significantly higher accuracy, with the r2-score improving from approximately 0.30 to 0.93 on average. The Extra Trees Regressor (ET) emerged as the best-performing model, demonstrating superior predictive capabilities across all datasets. These results underscore the potential of combining machine learning models with additional topological features from spectral analysis, offering deeper insights into the complex dynamics of solar activity and enhancing the precision of sunspot number predictions. This approach provides a novel methodology for improving space weather forecasting and contributes to a more comprehensive understanding of solar-terrestrial interactions.

Filament eruption by multiple reconnections

Context. Filament eruption is a common phenomenon in solar activity, but the triggering mechanism is not well understood. Aims. We focus our study on a filament eruption located in a complex nest of three active regions close to a coronal hole. Methods. The filament eruption is observed at multiple wavelengths: by the Global Oscillation Network Group (GONG), the Solar Terrestrial Relations Observatory (STEREO), the Solar Upper Transition Region Imager (SUTRI), and the Atmospheric Imaging Assembly (AIA) and Helioseismic and Magnetic Imager (HMI) on board the Solar Dynamic Observatory (SDO). Thanks to high-temporal-resolution observations, we were able to analyze the evolution of the fine structure of the filament in detail. The filament changes direction during the eruption, which is followed by a halo coronal mass ejection detected by the Large Angle Spectrometric Coronagraph (LASCO) on board the Solar and Heliospheric Observatory (SOHO). A Type III radio burst was also registered at the time of the eruption. To investigate the process of the eruption, we analyzed the magnetic topology of the filament region adopting a nonlinear force-free-field (NLFFF) extrapolation method and the polytropic global magnetohydrodynamic (MHD) modeling. We modeled the filament by embedding a twisted flux rope with the regularized Biot-Savart Laws (RBSL) method in the ambient magnetic field. Results. The extrapolation results show that magnetic reconnection occurs in a fan-spine configuration resulting in a circular flare ribbon. The global modeling of the corona demonstrates that there was an interaction between the filament and open field lines, causing a deflection of the filament in the direction of the observed CME eruption and dimming area. Conclusions. The modeling supports the following scenario: magnetic reconnection not only occurs with the filament itself (the flux rope) but also with the background magnetic field lines and open field lines of the coronal hole located to the east of the flux rope. This multiwavelength analysis indicates that the filament undergoes multiple magnetic reconnections on small and large scales with a drifting of the flux rope.

Breaking Boundaries: A Universal Wavefront Reconstruction Approach for High-resolution Solar Imaging

This Letter proposes a universal wavefront reconstruction approach based on a coupled data set and neural network, aiming to overcome the limitations of current algorithms in terms of universality and wavefront sensing accuracy for variable imaging objects. First, a novel data set, Multi-Object Wavefront Coupling Dataset (MOCD-Dataset), is developed to provide diverse data and enable the network to learn universal wavefront features. Next, a new universal wavefront reconstruction network called Object-Independent Wavefront Decoupling Network (OIWD-Net) is introduced, aiming to separate imaging object information from multiple variable images. Our algorithm eliminates the need for specialized wavefront sensors, has a simple system, high light energy utilization, and does not require customized models for each different type of imaging objects, making it highly practical. By combining the MOCD-Dataset and the OIWD-Net, excellent accuracy in wavefront reconstruction of different imaging objects has been achieved. This research provides a new solution for high-resolution image restoration in fields such as solar structure observation and astronomical high-resolution imaging.

AstroDLLC: Efficiently Reducing Storage and Transmission Costs for Massive Solar Observation Data via Deep Learning-based Lossless Compression

Effective data compression technology is essential for addressing data storage and transmission needs, especially given the escalating volume and complexity of data generated by contemporary astronomy. In this study, we propose utilizing deep learning-based lossless compression techniques to improve compression efficiency. We begin with a qualitative and quantitative analysis of the temporal and spatial redundancy in solar observation data. Based on this analysis, we introduce a novel deep learning-based framework called AstroDLLC for the lossless compression of astronomical solar images. AstroDLLC first segments high-resolution images into blocks to ensure that deep learning model training does not rely on high-computation power devices. It then addresses the non-normality of the partitioned data through simple reversible computational methods. Finally, it utilizes Bit-swap to train deep learning models that capture redundant features across multiple image frames, thereby enhancing compression efficiency. Comprehensive evaluations using data from the New Vacuum Solar Telescope reveal that AstroDLLC achieves a maximum compression ratio of 3.00 per image, surpassing Gzip, RICE, and other lossless technologies. The performance of AstroDLLC underscores its potential to address data compression challenges in astronomy.

The Effects of Pointing Error Sources on Energy Delivery from Orbiting Solar Reflectors

Many proposals are being made for cleaner, more sustainable forms of energy production. Terrestrial solar photovoltaic farms (SPFs) could potentially deliver large quantities of energy to the grid, although these are limited to daytime use. The output from these SPFs could be enhanced, particularly around dawn and dusk, by the use of orbiting solar reflectors (OSRs) in near-polar orbit. These would reflect an image of the solar disk, or solar image (SI), onto the SPFs to augment their energy output. Pointing requirements are therefore to ensure that reflected sunlight is delivered to the terrestrial SPF, avoiding the losses incurred by an offset of the SI and the SPF itself. The SI would typically be of order 10 km for a reflector in a 1000 km orbit. Given the potentially large size of the reflectors, this presents a challenge for the attitude determination and control system (ADCS) to ensure that the maximum quantity of energy can be delivered to a SPF, typically requiring large control moment gyro actuators. In addition, there exist numerous sources of error in the ADCS which can cause further degradation in the quantity of energy delivered to the SPF. These errors can manifest in the resolution of the various sensors, flexible structural modes, manufacturing inaccuracies, and misalignments due to vibration during launch. This paper will investigate the effects of pointing error sources (PES) on the reflector ADCS and so on the quantity of energy delivered to the SPF. With the application of a PD controller with feedforward compensation, and the set of noise characteristics defined in this paper, numerical simulations will show the typical losses in energy delivered to SPFs of 0.015% when the model accounts for PES in onboard sensors, actuator uncertainty and flexible structural modes.

Classification of Solar Cells EL Images with Different Busbars Via Deep Learning Models

Electricity generation from renewable energy sources such as solar energy has come to the forefront in the last decade. The solar energy cell is an indispensable part of the solar energy ecosystem of solar panels, and defective cells cause financial losses in energy production. Experienced experts are needed to detect defects on solar cells. Autonomous systems are important to accelerate the process. Classical image processing techniques are used to manually detect defects on cells. To use these techniques, many parameters are need to be entered into EL imaging software. However, in this study, these processes were carried out automatically without the need for external intervention. False detection/classification may occur during the processes performed by EL imaging devices due to weakness of the operator experience or EL imaging software. It is aimed to use automatic image processing and then deep learning techniques to achieve faster and higher performance than the results obtained from EL imaging devices using classic image processing techniques. AI algorithm and deep learning models can be an important solution. In this study, two AI algorithm and 10 different deep learning models were used to classify solar cells. EL images of defective and normal solar cells with 4 and 5 busbars were used in the study. The dataset, includes 9360 images of solar cells, 4680 of which are defective and 4680 are normal. Performance evaluation of the models made according to the confusion matrix. According to the results, Mobilenet-v2 and VGG-19 achieved the highest validation accuracy rate of 99.68%. According to F1-score, Mobilenetv2 achieved the highest performance of 99.73%. It has been shown that the Mobilenet-v2 is slightly more successful than other models in terms of validation and F1-score. The results show that trained DL models can be used as an inspection method in the production line of solar panels and cells.

Spectropolarimetric Radio Imaging of Faint Gyrosynchrotron Emission from a CME: A Possible Indication of the Insufficiency of Homogeneous Models

The geo-effectiveness of coronal mass ejections (CMEs) is determined primarily by their magnetic fields. Modeling of gyrosynchrotron (GS) emission is a promising remote sensing technique to measure the CME magnetic field at coronal heights. However, faint GS emission from CME flux ropes is hard to detect in the presence of bright solar emission from the solar corona. With high dynamic-range spectropolarimetric meter wavelength solar images provided by the Murchison Widefield Array, we have detected faint GS emission from a CME out to ∼8.3 R ⊙, the largest heliocentric distance reported to date. High-fidelity polarimetric calibration also allowed us to robustly detect circularly polarized emission from GS emission. For the first time in the literature, Stokes V detection has jointly been used with Stokes I spectra to constrain GS models. One expects that the inclusion of polarimetric measurement will provide tighter constraints on the GS model parameters. Instead, we found that homogeneous GS models, which have been used in all prior works, are unable to model both the total intensity and circular polarized emission simultaneously. This strongly suggests the need for using inhomogeneous GS models to robustly estimate the CME magnetic field and plasma parameters.

Deep Learning-Based Detection and Segmentation of Damage in Solar Panels

Renewable energy can lead to a sustainable future and solar energy is one the primary sources of renewable energy. Solar energy is harvested mainly by photovoltaic plants. Though there are a large number of solar panels, the economic efficiency of solar panels is not that high in comparison to energy production from coal or nuclear matter. The main risk involved in solar plants is the high maintenance cost involved in maintaining the plants. To help reduce this issue, automated solutions using Unmanned Aerial Vehicles (UAVs) and satellite imagery are proposed. In this research work, we propose a novel deep learning architecture for the segmentation of solar plant aerial images, which not only helps in automated solar plant maintenance, but can also be used for the area estimation and extraction of solar panels from an image. Along with this, we also propose a transfer learning-based model for the efficient classification of solar panel damage. Solar panel damage classification has a lot of applications. It can be integrated into monitoring systems, raising alerts when there is severe damage or damage of a certain type. The adaptive UNet model with Atrous Spatial Pyramid Pooling (ASPP) module that performed the dilated convolutions that we proposed achieved an overall accuracy of 98% with a Mean Intersection-Over-Union (IoU) Score of 95% and took under a second to process an image. Our classification model using Visual Geometry Group 19 (VGG19) as the backbone for feature extraction has achieved a classification accuracy of 98% with an F1 score of 99%, thus detecting the five classes of damage, including undamaged solar panels, in an efficient manner.

Solar multiobject multiframe blind deconvolution with a spatially variant convolution neural emulator

The study of astronomical phenomena through ground-based observations is always challenged by the distorting effects of Earth's atmosphere. Traditional methods of post facto image correction, essential for correcting these distortions, often rely on simplifying assumptions that limit their effectiveness, particularly in the presence of spatially variant atmospheric turbulence. Such cases are often solved by partitioning the field of view into small patches, deconvolving each patch independently, and merging all patches together. This approach is often inefficient and can produce artifacts. Recent advancements in computational techniques and the advent of deep learning offer new pathways to address these limitations. This paper introduces a novel framework leveraging a deep neural network to emulate spatially variant convolutions, offering a breakthrough in the efficiency and accuracy of astronomical image deconvolution. By training on a dataset of images convolved with spatially invariant point spread functions and validating its generalizability to spatially variant conditions, this approach presents a significant advancement over traditional methods. The convolution emulator is used as a forward model in a multiobject multiframe blind deconvolution algorithm for solar images. The emulator enables the deconvolution of solar observations across large fields of view without resorting to patch-wise mosaicking, thus avoiding the artifacts associated with such techniques. This method represents a significant computational advantage, reducing processing times by orders of magnitude.