Recent advancements in environmental monitoring are increasingly focused on enhancing the accuracy and efficiency of pollution detection and forecasting. For instance, new frameworks like OilSAM2 leverage memory-augmented segmentation to improve oil spill detection from Synthetic Aperture Radar imagery, addressing challenges posed by appearance variability and temporal coherence. Similarly, the introduction of 6ABOS automates atmospheric corrections for hyperspectral imagery, facilitating better water quality assessments. In indoor environments, novel forecasting models are being developed to predict air quality fluctuations, integrating human activity data to improve responsiveness. Long-horizon wildfire risk forecasting is also evolving, with hierarchical diffusion models streamlining predictions over extended periods. Additionally, innovative approaches like SPyCer utilize satellite imagery for near-surface air temperature estimation, while benchmarking studies on coastal hypoxia forecasting are refining machine learning techniques for real-time ecosystem management. Collectively, these efforts aim to provide more precise and actionable insights for environmental challenges, enhancing both ecological and public health outcomes.
Segmenting oil spills from Synthetic Aperture Radar (SAR) imagery remains challenging due to severe appearance variability, scale heterogeneity, and the absence of temporal continuity in real world mo...
Indoor air quality (IAQ) forecasting plays a critical role in safeguarding occupant health, ensuring thermal comfort, and supporting intelligent building control. However, predicting future concentrat...
The Environmental Mapping and Analysis Program (EnMAP) mission has opened new frontiers in the monitoring of optically complex environments. However, the accurate retrieval of surface reflectance over...
Long-horizon wildfire risk forecasting requires generating probabilistic spatial fields under sparse event supervision while maintaining computational efficiency across multiple prediction horizons. E...
Modern Earth observation relies on satellites to capture detailed surface properties. Yet, many phenomena that affect humans and ecosystems unfold in the atmosphere close to the surface. Near-ground s...
Coastal hypoxia, especially in the northern part of Gulf of Mexico, presents a persistent ecological and economic concern. Seasonal models offer coarse forecasts that miss the fine-scale variability n...
Vision Transformers have achieved remarkable success in spatio-temporal prediction, but their scalability remains limited for ultra-high-resolution, continent-scale domains required in real-world envi...