Vít Růžička

We introduce the Differential Matched Filter (DiffMF), a computationally efficient extension to the Matched Filter (MF) designed to distinguish genuine Greenhouse Gas (GHG) enhancements from false enhancements. Conventional MF retrievals are often plagued by high-magnitude false concentrations driven by surface or atmospheric features that mimic the target absorption. DiffMF exploits the inherent functional characteristics of wavelength-varying spectroscopic observations—specifically, the derivatives of the whitened target and observed spectrum with respect to wavelength—to address the numerical ambiguities at the source of these artifacts. We validate this approach on a dataset consisting of 900 EMIT scenes containing 389 expert labeled CH₄ plumes and 5,763 "plume-like" false plume detections, and demonstrate that simple metrics derived from our DiffMF retrievals provide an effective means to discriminate real GHG-enhanced pixels from false enhancements, to reject of false positive plume detections, and to guide iterative refinement of background covariance estimates without compromising the broadly effective filtering capabilities provided by the conventional MF. We further present preliminary results extending DiffMF to the retrieval of other GHG species including CO₂, NO₂, and NH₃. To facilitate adoption across imaging spectroscopy instruments, platforms, and applications, we provide a complete Python implementation for reproducibility.

We introduce the Differential Matched Filter (DiffMF), a computationally efficient extension to the Matched Filter (MF) designed to distinguish genuine Greenhouse Gas (GHG) enhancements from false enhancements. Conventional MF retrievals are often plagued by high-magnitude false concentrations driven by surface or atmospheric features that mimic the target absorption. DiffMF exploits the inherent functional characteristics of wavelength-varying spectroscopic observations—specifically, the derivatives of the whitened target and observed spectrum with respect to wavelength—to address the numerical ambiguities at the source of these artifacts. We validate this approach on a dataset consisting of 900 EMIT scenes containing 389 expert labeled CH₄ plumes and 5,763 "plume-like" false plume detections, and demonstrate that simple metrics derived from our DiffMF retrievals provide an effective means to discriminate real GHG-enhanced pixels from false enhancements, to reject of false positive plume detections, and to guide iterative refinement of background covariance estimates without compromising the broadly effective filtering capabilities provided by the conventional MF. We further present preliminary results extending DiffMF to the retrieval of other GHG species including CO₂, NO₂, and NH₃. To facilitate adoption across imaging spectroscopy instruments, platforms, and applications, we provide a complete Python implementation for reproducibility.

We introduce the Differential Matched Filter (DiffMF), a computationally efficient extension to the Matched Filter (MF) designed to distinguish genuine Greenhouse Gas (GHG) enhancements from false enhancements. Conventional MF retrievals are often plagued by high-magnitude false concentrations driven by surface or atmospheric features that mimic the target absorption. DiffMF exploits the inherent functional characteristics of wavelength-varying spectroscopic observations—specifically, the derivatives of the whitened target and observed spectrum with respect to wavelength—to address the numerical ambiguities at the source of these artifacts. We validate this approach on a dataset consisting of 900 EMIT scenes containing 389 expert labeled CH₄ plumes and 5,763 "plume-like" false plume detections, and demonstrate that simple metrics derived from our DiffMF retrievals provide an effective means to discriminate real GHG-enhanced pixels from false enhancements, to reject of false positive plume detections, and to guide iterative refinement of background covariance estimates without compromising the broadly effective filtering capabilities provided by the conventional MF. We further present preliminary results extending DiffMF to the retrieval of other GHG species including CO₂, NO₂, and NH₃. To facilitate adoption across imaging spectroscopy instruments, platforms, and applications, we provide a complete Python implementation for reproducibility.

We introduce the Differential Matched Filter (DiffMF), a computationally efficient extension to the Matched Filter (MF) designed to distinguish genuine Greenhouse Gas (GHG) enhancements from false enhancements. Conventional MF retrievals are often plagued by high-magnitude false concentrations driven by surface or atmospheric features that mimic the target absorption. DiffMF exploits the inherent functional characteristics of wavelength-varying spectroscopic observations—specifically, the derivatives of the whitened target and observed spectrum with respect to wavelength—to address the numerical ambiguities at the source of these artifacts. We validate this approach on a dataset consisting of 900 EMIT scenes containing 389 expert labeled CH₄ plumes and 5,763 "plume-like" false plume detections, and demonstrate that simple metrics derived from our DiffMF retrievals provide an effective means to discriminate real GHG-enhanced pixels from false enhancements, to reject of false positive plume detections, and to guide iterative refinement of background covariance estimates without compromising the broadly effective filtering capabilities provided by the conventional MF. We further present preliminary results extending DiffMF to the retrieval of other GHG species including CO₂, NO₂, and NH₃. To facilitate adoption across imaging spectroscopy instruments, platforms, and applications, we provide a complete Python implementation for reproducibility.

We introduce the Differential Matched Filter (DiffMF), a computationally efficient extension to the Matched Filter (MF) designed to distinguish genuine Greenhouse Gas (GHG) enhancements from false enhancements. Conventional MF retrievals are often plagued by high-magnitude false concentrations driven by surface or atmospheric features that mimic the target absorption. DiffMF exploits the inherent functional characteristics of wavelength-varying spectroscopic observations—specifically, the derivatives of the whitened target and observed spectrum with respect to wavelength—to address the numerical ambiguities at the source of these artifacts. We validate this approach on a dataset consisting of 900 EMIT scenes containing 389 expert labeled CH4 plumes and 5,763 "plume-like" false plume detections, and demonstrate that simple metrics derived from our DiffMF retrievals provide an effective means to discriminate real GHG-enhanced pixels from false enhancements, to reject of false positive plume detections, and to guide iterative refinement of background covariance estimates without compromising the broadly effective filtering capabilities provided by the conventional MF. We further present preliminary results extending DiffMF to the retrieval of other GHG species including CO2, NO2, and NH3. To facilitate adoption across imaging spectroscopy instruments, platforms, and applications, we provide a complete Python implementation for reproducibility.

Coronagraphic observations enable direct monitoring of coronal mass ejections (CMEs) through scattered light from free electrons, but determining the 3D plasma distribution from 2D imaging data is challenging due to the optically thin plasma and the complex image formation. We introduce Sun Neural Radiance Field for CMEs (SuNeRF-CME), a framework for 3D tomographic reconstructions of the heliosphere using multiviewpoint coronagraphic observations. The method uses a neural radiance-field to estimate the electron density in the heliosphere through ray tracing, while accounting for the underlying Thomson scattering. The model is optimized by iteratively fitting the time-dependent observational data. In addition, we apply physical constraints in terms of continuity, propagation direction, and speed of the heliospheric plasma to overcome limitations imposed by the sparse number of viewpoints. We utilize synthetic observations of a CME simulation to quantify the model’s performance for different viewpoint configurations. Within this controlled synthetic setting, the results demonstrate that our method can reliably estimate the CME parameters from only two viewpoints, with a mean velocity error of 3.01% ± 1.94% and propagation direction errors of 3 . ° 39 ± 1 . ° 94 in latitude and 1 . ° 76 ± 0 . ° 79 in longitude. We further show that our approach can achieve a 3D reconstruction of the simulated CME from two viewpoints, where we correctly model the three-part structure, deformed CME front, and internal plasma variations. Additional viewpoints can be seamlessly integrated, directly enhancing the reconstruction of the plasma distribution in the heliosphere. These results demonstrate the potential of physics-informed radiance-field methods for CME tomography, paving the way for future extensions toward observational data and space weather applications.

We present results from deploying machine learning models into production for detecting methane leaks in data from imaging spectroscopy satellites. We use and publicly release one of the largest global and most diverse datasets of methane plumes from three different imaging spectroscopy missions. We compute three different variants of methane enhancement products and quantitatively compare their suitability for using with deep learning models (showing an improvement of more than 47% in the F1 score). Using the data from different sensors, we demonstrate generalisation capabilities and transfer knowledge learned on the largest dataset to the other smaller sets. With attention to the capabilities needed for model deployment, we extend prior evaluation methodologies from small tiled datasets to full granule evaluation. Using these, we identify that most typically used methods create a large number of false alerts. We propose model ensembles, which reduce the number of false alarms by over 57%, while further improving the F1 score by additional 3.2% over the non ensembled variant. Finally, we deploy the proposed models in a real methane leak detection pipeline used at the United Nations Environment Programme’s Methane Alert and Response System. Detections of our model are used as proposals for a team of analysts, which effectively speeds up the detection process and reduces the time needed to make these detections on each capture. In 9 months our system helped detect over 1465 methane leak events in data from EMIT, PRISMA and EnMAP. The created datasets are part of a living system and ready to be used for model adaptation and improvement over time. Our work is an important step towards a global AI assisted methane leak detection system, which will be necessary in the light of a large number of new imaging spectroscopy missions being launched by NASA, ESA and private actors. Relying solely on manual data detection is no longer feasible.

On-board processing of hyperspectral data with machine learning models would enable an unprecedented amount of autonomy across a wide range of tasks allowing new capabilities, such as early warning systems and automated scheduling across constellations, of satellites. However, current classical methods suffer from high false positive rates and therefore prevent easy automation while previously published deep learning models exhibit prohibitive computational requirements. We propose fast and accurate machine learning architectures, which support end-to-end processing of data with high spectral dimension without relying on handcrafted products or spectral band compression techniques. We create three new large datasets of hyperspectral data containing all relevant spectral bands from the near global sensor EMIT. We evaluate our models on two tasks related to hyperspectral data processing—methane detection and mineral identification. Our models reach a new state-of-the-art performance on the task of methane detection, where we improve the F1 score of previous deep learning models by 27% on a newly created synthetic dataset and by 13% on the previously released large benchmark dataset. Our models generalize from synthetic datasets to data with real methane leak events and boost performance by 6.9% in F1 score in contrast with training models from scratch on the real data. Finally, with our newly proposed architectures, one capture from the EMIT sensor can be processed within 30 s on a realistic proxy of the ION-SCV 004 satellite and in less than 0.64 s on a GPU powered Jetson AGX Xavier board.

Nano and microsatellites have expanded the acquisition of satellite images with higher spatial, temporal, and spectral resolutions. Nevertheless, downlinking all this data to the ground for processing becomes challenging as the amount of remote sensing data rises. Custom onboard algorithms are designed to make real-time decisions and to prioritize and reduce the amount of data transmitted to the ground. However, these onboard algorithms frequently require cloud-free bottom-of-atmosphere surface reflectance estimations as inputs to operate. In this context, this paper presents DTACSNet, an onboard cloud detection and atmospheric correction processor based on lightweight Convolutional Neural Networks. DTACSNet provides cloud and cloud shadow masks and surface reflectance estimates 10x faster than the operational Sentinel-2 L2A processor in dedicated space-tested hardware: 7 mins vs. 1 hour for a 10,980 x 10,980 scene. The DTACSNet cloud masking, based on a lightweight neural network, obtains the highest F2-score (0.81), followed by the state-of-the-art KappaMask (0.74), Fmask (0.72) and Sen2Cor v.2.8 (0.51) algorithms. Additionally, validation results on independent datasets show that DTACSNet can efficiently replicate Sen2Cor surface reflectance estimates, reporting a competitive accuracy with differences below 2%. Code is available at https://github.com/spaceml-org/DTACSNet .