Kélian Sommer
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Kélian Sommer has a total of 42 co-authors. Here are the top co-authors by number of shared publications:
- Bertrand Plez
LUPM Univ Montpellier CNRS Montpellier France
6 shared publications - Romain Brunet
Aix-Marseille Université, CNRS, CNES, LAM, Marseille, France
6 shared publications - Wassim Kabalan
Université Paris Cité, CNRS, Astroparticule et Cosmologie, F-75013 Paris, France
6 shared publications - Jeremy Neveu
Laboratoire de Physique des deux infinis Irène Joliot-Curie (IJCLab), Université Paris-Saclay, CNRS/IN2P3
5 shared publications - Johann Cohen-Tanugi
Université Clermont-Auvergne, CNRS, LPCA
5 shared publications - Laurent Le Guillou
Sorbonne Universite
5 shared publications
Latest Publications
- 8 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Photométrie de haute précision pour les grands relevés cosmologiques ZTF et LSST : caractérisation des transmissions instrumentales et atmosphériques.
Les supernovae de type 1a (SNe Ia) sont des sondes cosmologiques essentielles pour étudier l'Univers récent et approfondir notre compréhension de l'énergie noire. Leur luminosité régulière, combinée aux observations réalisées par les grands relevés photométriques terrestres multi-bandes, permet d'estimer leurs distances et de contraindre les paramètres cosmologiques. Toutefois, pour les relevés actuels et futurs comme le textit{Zwicky Transient Facility} (ZTF) et le textit{Legacy Survey of Space and Time} (LSST), les incertitudes de calibration photométrique surpasseront les incertitudes statistiques dans les analyses cosmologiques basées sur les SNe Ia, ce qui limitera la précision des résultats cosmologiques. Pour atteindre les objectifs fixés par la collaboration textit{Dark Energy Science Collaboration} du relevé textit{LSST}, il est donc nécessaire d'améliorer la précision de la calibration des flux photométriques d'un ordre de grandeur, jusqu'à 0.1%.Pour obtenir des rapports de distance précis entre les observations de SNe Ia à différents textit{redshifts}, une calibration rigoureuse des couleurs est essentielle, car toute erreur dans la réponse chromatique des instruments peut fausser notre compréhension de l'expansion de l'univers. Par conséquent, une connaissance précise de la transmission totale, incluant les effets instrumentaux et atmosphériques à chaque époque du relevé, est indispensable pour atteindre le niveau de précision requis.Les travaux menés au cours de cette thèse se concentrent sur le développement de nouvelles méthodes pour améliorer la caractérisation de ces deux fonctions de transmission, formant ainsi deux axes de recherche complémentaires intégrés au projet de métrologie StarDICE qui vise à relever le défi de la calibration photométrique en utilisant des sources lumineuses calibrées en laboratoire pour étendre la calibration absolue à un catalogue d'étoiles standards, observées par les télescopes terrestres participant à des relevés cosmologiques.Le premier axe de travail concerne la conception d'une version transportable du textit{Collimated Beam Projector}, un instrument de laboratoire destiné à déterminer avec précision les réponses spectrales des télescopes effectuant des relevés photométriques au sol. J'ai contribué à la mise au point d'un prototype sur banc expérimental, démontrant qu'il est possible d'atteindre une précision de l'ordre d'une millimagnitude sur l'intégrale des fonctions de transmission instrumentales des filtres textit{g} et textit{r} du système photométrique de ZTF.Le second axe repose sur l'exploration d'une méthode inédite de détermination de l'extinction grise atmosphérique par des mesures radiométriques fournies par une caméra infrarouge thermique. Déployée avec succès à l'Observatoire de Haute-Provence, cette approche s'avère prometteuse pour qualifier la structure spatiale de l'extinction grise lors de conditions d'observation variables et corriger les mesures photométriques avec une précision supérieures aux méthodes déjà existantes.Ces deux efforts ont montré des résultats encourageants pour les développements futurs, tant pour la mesure de la réponse instrumentale du télescope du relevé ZTF que pour la caractérisation de l'extinction atmosphérique grise du relevé LSST qui mènera sa campagne d'observation même en conditions atmosphériques suboptimales.
- 8 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
Photométrie de haute précision pour les grands relevés cosmologiques ZTF et LSST : caractérisation des transmissions instrumentales et atmosphériques.
Les supernovae de type 1a (SNe Ia) sont des sondes cosmologiques essentielles pour étudier l'Univers récent et approfondir notre compréhension de l'énergie noire. Leur luminosité régulière, combinée aux observations réalisées par les grands relevés photométriques terrestres multi-bandes, permet d'estimer leurs distances et de contraindre les paramètres cosmologiques. Toutefois, pour les relevés actuels et futurs comme le textit{Zwicky Transient Facility} (ZTF) et le textit{Legacy Survey of Space and Time} (LSST), les incertitudes de calibration photométrique surpasseront les incertitudes statistiques dans les analyses cosmologiques basées sur les SNe Ia, ce qui limitera la précision des résultats cosmologiques. Pour atteindre les objectifs fixés par la collaboration textit{Dark Energy Science Collaboration} du relevé textit{LSST}, il est donc nécessaire d'améliorer la précision de la calibration des flux photométriques d'un ordre de grandeur, jusqu'à 0.1%.Pour obtenir des rapports de distance précis entre les observations de SNe Ia à différents textit{redshifts}, une calibration rigoureuse des couleurs est essentielle, car toute erreur dans la réponse chromatique des instruments peut fausser notre compréhension de l'expansion de l'univers. Par conséquent, une connaissance précise de la transmission totale, incluant les effets instrumentaux et atmosphériques à chaque époque du relevé, est indispensable pour atteindre le niveau de précision requis.Les travaux menés au cours de cette thèse se concentrent sur le développement de nouvelles méthodes pour améliorer la caractérisation de ces deux fonctions de transmission, formant ainsi deux axes de recherche complémentaires intégrés au projet de métrologie StarDICE qui vise à relever le défi de la calibration photométrique en utilisant des sources lumineuses calibrées en laboratoire pour étendre la calibration absolue à un catalogue d'étoiles standards, observées par les télescopes terrestres participant à des relevés cosmologiques.Le premier axe de travail concerne la conception d'une version transportable du textit{Collimated Beam Projector}, un instrument de laboratoire destiné à déterminer avec précision les réponses spectrales des télescopes effectuant des relevés photométriques au sol. J'ai contribué à la mise au point d'un prototype sur banc expérimental, démontrant qu'il est possible d'atteindre une précision de l'ordre d'une millimagnitude sur l'intégrale des fonctions de transmission instrumentales des filtres textit{g} et textit{r} du système photométrique de ZTF.Le second axe repose sur l'exploration d'une méthode inédite de détermination de l'extinction grise atmosphérique par des mesures radiométriques fournies par une caméra infrarouge thermique. Déployée avec succès à l'Observatoire de Haute-Provence, cette approche s'avère prometteuse pour qualifier la structure spatiale de l'extinction grise lors de conditions d'observation variables et corriger les mesures photométriques avec une précision supérieures aux méthodes déjà existantes.Ces deux efforts ont montré des résultats encourageants pour les développements futurs, tant pour la mesure de la réponse instrumentale du télescope du relevé ZTF que pour la caractérisation de l'extinction atmosphérique grise du relevé LSST qui mènera sa campagne d'observation même en conditions atmosphériques suboptimales.
- Open Access
- 6 January 2026
- Published by Oxford University Press (OUP) in RAS Techniques and Instruments
- Vol. 5
StarDICE IV: correcting visible photometry from atmospheric grey extinction using thermal infrared observations
Ground-based surveys such as the Vera C. Rubin Observatory’s Legacy Survey of Space and Time require photometric calibration that is both long-term stable and spatially uniform at the sub-per cent level, even during non-photometric conditions. Achieving this precision motivates new approaches to characterize atmospheric transmission, particularly to mitigate grey extinction from clouds. The StarDICE experiment aims to establish a metrology chain linking laboratory standards to astrophysical fluxes with 1 mmag accuracy in the griz bands, a goal for which controlling variable atmospheric effects is essential. We present a method that corrects photometric measurements using simultaneous radiometric information from an infrared thermal camera. The grey-extinction model is fit on an image-by-image basis using thermal radiance excess and the difference between synthetic and instrumental fluxes of calibration stars, without requiring assumptions about the spatial structure of extinction. The method relies on a forward model that incorporates environmental monitoring, radiative-transfer simulations, and Gaia DR3 stellar catalogues. Using data from a remote observing system that repeatedly monitored two fields under diverse atmospheric conditions, we show that the corrections reduce residuals between corrected and reference magnitudes and produce extinction maps with 2-arcmin resolution and $\sim$0.01 mag accuracy. Using this technique, we can recover data acquired under non-photometric conditions with a precision comparable to data obtained under photometric conditions. For the most affected exposures, the mean absolute error improves from 0.64 to 0.11 mag, and temporal extinction variations can be reduced to 0.025 mag per source. We discuss the implications of this technique for future surveys and outline directions for further refinement.
- 6 January 2026
- Published by Oxford University Press (OUP) in RAS Techniques and Instruments
- Vol. 5
Open AccessStarDICE IV: correcting visible photometry from atmospheric grey extinction using thermal infrared observations
Ground-based surveys such as the Vera C. Rubin Observatory’s Legacy Survey of Space and Time require photometric calibration that is both long-term stable and spatially uniform at the sub-per cent level, even during non-photometric conditions. Achieving this precision motivates new approaches to characterize atmospheric transmission, particularly to mitigate grey extinction from clouds. The StarDICE experiment aims to establish a metrology chain linking laboratory standards to astrophysical fluxes with 1 mmag accuracy in the griz bands, a goal for which controlling variable atmospheric effects is essential. We present a method that corrects photometric measurements using simultaneous radiometric information from an infrared thermal camera. The grey-extinction model is fit on an image-by-image basis using thermal radiance excess and the difference between synthetic and instrumental fluxes of calibration stars, without requiring assumptions about the spatial structure of extinction. The method relies on a forward model that incorporates environmental monitoring, radiative-transfer simulations, and Gaia DR3 stellar catalogues. Using data from a remote observing system that repeatedly monitored two fields under diverse atmospheric conditions, we show that the corrections reduce residuals between corrected and reference magnitudes and produce extinction maps with 2-arcmin resolution and $\sim$0.01 mag accuracy. Using this technique, we can recover data acquired under non-photometric conditions with a precision comparable to data obtained under photometric conditions. For the most affected exposures, the mean absolute error improves from 0.64 to 0.11 mag, and temporal extinction variations can be reduced to 0.025 mag per source. We discuss the implications of this technique for future surveys and outline directions for further refinement.
- Open Access
- 12 May 2025
- Published by Copernicus GmbH in Atmospheric Measurement Techniques
- Vol. 18 (9) , 2083-2101
Infrared radiometric image classification and segmentation of cloud structures using a deep-learning framework from ground-based infrared thermal camera observations
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. The precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structures in gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with convolutional neural networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations based on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 95 % in our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- 12 May 2025
- Published by Copernicus GmbH in Atmospheric Measurement Techniques
- Vol. 18 (9) , 2083-2101
Open AccessInfrared radiometric image classification and segmentation of cloud structures using a deep-learning framework from ground-based infrared thermal camera observations
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. The precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structures in gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with convolutional neural networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations based on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 95 % in our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- Open Access
- 3 January 2025
- Published by Oxford University Press (OUP) in RAS Techniques and Instruments
- Vol. 4
StarDICE III: characterization of the photometric instrument with a collimated beam projector
The measurement of Type Ia supernovae magnitudes provides cosmological distances, which constrain dark energy parameters. Current and upcoming large photometric surveys require improved photometric calibration precision to reduce systematic uncertainties in cosmological constraints. The StarDICE experiment aims to establish accurate broad-band flux references for these surveys, targeting sub-percent precision in magnitude measurements. Achieving this requires precise filter bandpass measurements for both StarDICE and survey instruments with sub-nanometre accuracy. To this end, we developed the Collimated Beam Projector (CBP), an optical device for calibrating the throughput of astronomical telescopes and their filters. The CBP uses a tunable laser source and a reversed telescope to emit a parallel monochromatic light beam, continuously monitored in flux and wavelength. The CBP output flux is measured with a large-area photodiode calibrated relative to a NIST photodiode. Using CBP measurements, we derive the StarDICE telescope throughput and filter transmissions, anchoring them to NIST’s absolute calibration. After analysing systematic uncertainties, we achieved sub-nanometre accuracy for filter central wavelengths, measured filter transmission with ~0.5 per cent precision per $1 \,{\rm nm}$ bin, and detected out-of-band leakages at a relative level of $10^{-4}$. Furthermore, we synthesized equivalent transmission for full pupil illumination from four sampled positions in the StarDICE telescope mirror, with ~$0.2 \,{\rm nm}$ accuracy for central wavelengths and $7 \,{\rm mmag}$ for broad-band fluxes. This demonstrates our ability to characterize telescope throughput down to the millimagnitude, paving the way for future developments, such as the Rubin-CBP for measuring the LSST at Vera Rubin Observatory, and a portable CBP version for in-situ transmission monitoring.- 3 January 2025
- Published by Oxford University Press (OUP) in RAS Techniques and Instruments
- Vol. 4
Open AccessStarDICE III: characterization of the photometric instrument with a collimated beam projector
The measurement of Type Ia supernovae magnitudes provides cosmological distances, which constrain dark energy parameters. Current and upcoming large photometric surveys require improved photometric calibration precision to reduce systematic uncertainties in cosmological constraints. The StarDICE experiment aims to establish accurate broad-band flux references for these surveys, targeting sub-percent precision in magnitude measurements. Achieving this requires precise filter bandpass measurements for both StarDICE and survey instruments with sub-nanometre accuracy. To this end, we developed the Collimated Beam Projector (CBP), an optical device for calibrating the throughput of astronomical telescopes and their filters. The CBP uses a tunable laser source and a reversed telescope to emit a parallel monochromatic light beam, continuously monitored in flux and wavelength. The CBP output flux is measured with a large-area photodiode calibrated relative to a NIST photodiode. Using CBP measurements, we derive the StarDICE telescope throughput and filter transmissions, anchoring them to NIST’s absolute calibration. After analysing systematic uncertainties, we achieved sub-nanometre accuracy for filter central wavelengths, measured filter transmission with ~0.5 per cent precision per $1 \,{\rm nm}$ bin, and detected out-of-band leakages at a relative level of $10^{-4}$. Furthermore, we synthesized equivalent transmission for full pupil illumination from four sampled positions in the StarDICE telescope mirror, with ~$0.2 \,{\rm nm}$ accuracy for central wavelengths and $7 \,{\rm mmag}$ for broad-band fluxes. This demonstrates our ability to characterize telescope throughput down to the millimagnitude, paving the way for future developments, such as the Rubin-CBP for measuring the LSST at Vera Rubin Observatory, and a portable CBP version for in-situ transmission monitoring. - Open Access
- 28 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Reply on RC2
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- 28 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Open AccessReply on RC2
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- Open Access
- 28 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Reply on RC1
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- 28 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Open AccessReply on RC1
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- Open Access
- 12 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Comment on egusphere-2024-101
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- 12 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Open AccessComment on egusphere-2024-101
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- Open Access
StarDICE II: Calibration of an Uncooled Infrared Thermal Camera for Atmospheric Gray Extinction Characterization
The StarDICE experiment strives to establish an instrumental metrology chain with a targeted accuracy of 1 mmag in griz bandpasses to meet the calibration requirements of next-generation cosmological surveys. Atmospheric transmission is a significant source of systematic uncertainty. We propose a solution relying on an uncooled infrared thermal camera to evaluate gray extinction variations. However, achieving accurate measurements with thermal imaging systems necessitates prior calibration due to temperature-induced effects, compromising their spatial and temporal precision. Moreover, these systems cannot provide scene radiance in physical units by default. This study introduces a new calibration process utilizing a tailored forward modeling approach. The method incorporates sensor, housing, flat-field support, and ambient temperatures, along with raw digital response, as input data. Experimental measurements were conducted inside a climatic chamber, with a FLIR Tau2 camera imaging a thermoregulated blackbody source. The results demonstrate the calibration effectiveness, achieving precise radiance measurements with a temporal pixel dispersion of 0.09 and residual spatial noise of 0.03 . We emphasize that the accuracy of scene radiance retrieval can be systematically affected by the camera’s close thermal environment, especially when the ambient temperature exceeds that of the scene.Open AccessStarDICE II: Calibration of an Uncooled Infrared Thermal Camera for Atmospheric Gray Extinction Characterization
The StarDICE experiment strives to establish an instrumental metrology chain with a targeted accuracy of 1 mmag in griz bandpasses to meet the calibration requirements of next-generation cosmological surveys. Atmospheric transmission is a significant source of systematic uncertainty. We propose a solution relying on an uncooled infrared thermal camera to evaluate gray extinction variations. However, achieving accurate measurements with thermal imaging systems necessitates prior calibration due to temperature-induced effects, compromising their spatial and temporal precision. Moreover, these systems cannot provide scene radiance in physical units by default. This study introduces a new calibration process utilizing a tailored forward modeling approach. The method incorporates sensor, housing, flat-field support, and ambient temperatures, along with raw digital response, as input data. Experimental measurements were conducted inside a climatic chamber, with a FLIR Tau2 camera imaging a thermoregulated blackbody source. The results demonstrate the calibration effectiveness, achieving precise radiance measurements with a temporal pixel dispersion of 0.09 and residual spatial noise of 0.03 . We emphasize that the accuracy of scene radiance retrieval can be systematically affected by the camera’s close thermal environment, especially when the ambient temperature exceeds that of the scene. - Open Access
- 10 July 2024
- Published by MDPI AG in Preprints.org
StarDICE II: calibration of an uncooled infrared thermal camera for atmospheric gray extinction characterization
The StarDICE experiment strives to establish an instrumental metrology chain with a targeted accuracy of 1 mmag in griz bandpasses to meet the calibration requirements of next-generation cosmological surveys. Atmospheric transmission stands out as a significant source of systematic uncertainty. Specifically, gray extinction induces spurious variations of photometric fluxes. We propose a solution relying on an uncooled long-wave infrared thermal camera to evaluate it. This type of camera can provide real-time insights into atmospheric conditions and detect cirrus clouds. However, achieving accurate measurements with thermal imaging systems necessitates prior calibration due to the influence of temperature-induced effects, which compromise their spatial and temporal precision. Moreover, these systems cannot provide scene radiance values in physical units by default. This study introduces a new calibration process utilizing a tailored forward modeling approach. The method is applied to an uncooled FLIR Tau2 thermal camera. It incorporates sensor, housing, flat-field support, and ambient temperatures, along with raw digital response, as input data. Experimental measurements are conducted inside a climatic chamber, with the camera imaging a thermoregulated blackbody source. Results demonstrate the calibration effectiveness, achieving precise radiance measurements with a temporal pixel root-mean-square error (RMSE) of 0.09 W/m2/srand residual spatial noise of 0.03 W/m2/sr.
- 10 July 2024
- Published by MDPI AG in Preprints.org
Open AccessStarDICE II: calibration of an uncooled infrared thermal camera for atmospheric gray extinction characterization
The StarDICE experiment strives to establish an instrumental metrology chain with a targeted accuracy of 1 mmag in griz bandpasses to meet the calibration requirements of next-generation cosmological surveys. Atmospheric transmission stands out as a significant source of systematic uncertainty. Specifically, gray extinction induces spurious variations of photometric fluxes. We propose a solution relying on an uncooled long-wave infrared thermal camera to evaluate it. This type of camera can provide real-time insights into atmospheric conditions and detect cirrus clouds. However, achieving accurate measurements with thermal imaging systems necessitates prior calibration due to the influence of temperature-induced effects, which compromise their spatial and temporal precision. Moreover, these systems cannot provide scene radiance values in physical units by default. This study introduces a new calibration process utilizing a tailored forward modeling approach. The method is applied to an uncooled FLIR Tau2 thermal camera. It incorporates sensor, housing, flat-field support, and ambient temperatures, along with raw digital response, as input data. Experimental measurements are conducted inside a climatic chamber, with the camera imaging a thermoregulated blackbody source. Results demonstrate the calibration effectiveness, achieving precise radiance measurements with a temporal pixel root-mean-square error (RMSE) of 0.09 W/m2/srand residual spatial noise of 0.03 W/m2/sr.
- Open Access
- 4 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Comment on egusphere-2024-101
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.
- 4 July 2024
- Published by Copernicus GmbH
- Vol. 2024, 1-32
Open AccessComment on egusphere-2024-101
Infrared thermal cameras offer reliable means of assessing atmospheric conditions by measuring the downward radiance from the sky, facilitating their usage in cloud monitoring endeavors. Precise identification and detection of clouds in images pose great challenges stemming from the indistinct boundaries inherent to cloud formations. Various methodologies for segmentation have been previously suggested. Most of them rely on color as the distinguishing criterion for cloud identification in the visible spectral domain and thus lack the ability to detect cloud structure on gray-scaled images with satisfying accuracy. In this work, we propose a new complete deep-learning framework to perform image classification and segmentation with Convolutional Neural Networks. We demonstrate the effectiveness of this technique by conducting a series of tests and validations on self-captured infrared sky images. Our findings reveal that the models can effectively differentiate between image types and accurately capture detailed cloud structure information at the pixel level, even when trained with a single binary ground-truth mask per input sample. The classifier model achieves an excellent accuracy of 99 % in image type distinction, while the segmentation model attains a mean pixel accuracy of 94 % on our dataset. We emphasize that our framework exhibits strong viability and can be used for infrared thermal ground-based cloud monitoring operations over extended durations. We expect to take advantage of this framework for astronomical applications by providing cloud cover selection criteria for ground-based photometric observations within the StarDICE experiment.