Paper detects anomalies in wheat and rapeseed crops using satellite data.
problem Detecting anomalies in crop development at parcel-level.
method Unsupervised outlier detection using SAR and multispectral features.
result Best performance with a 10% outlier ratio, achieving 94.1% true positives for rapeseed and 95.5% for wheat.
Proposes a method to identify key components for predicting epidemic dynamics with limited resources.
problem Predicting epidemic dynamics with limited surveillance resources.
method Developed a group sparse Bayesian learning algorithm to identify sentinel components for monitoring.
result The proposed algorithm effectively predicts epidemic dynamics using partial data from sentinel components.
Forest tree species mapped with high accuracy using satellite data.
problem Classifying dominant tree species in Swedish forests.
method Extreme gradient boosting model with Bayesian optimization, combining Sentinel-1/2 satellite data and field observations.
result Overall accuracy of 85%, F1 score of 0.82, Matthews correlation coefficient of 0.81.
Big data trend has enforced the data-centric systems to have continuous fast data streams. In recent years, real-time analytics on stream data has formed into a new research field, which aims to answer queries about what-is-happening-now with a negligible delay. The real challenge with real-time stream data processing …
Sentinel improves time series forecasting by modeling both temporal and channel dependencies.
problem Limited effectiveness of existing transformer-based architectures in multivariate time-series forecasting.
method Proposes Sentinel, a full transformer-based architecture with multi-patch attention mechanism.
result Sentinel achieves better or comparable performance compared to state-of-the-art approaches.
Study uses SAR data to estimate forest vegetation indices, improving monitoring of temperate forests.
problem Limitations of optical satellite data in monitoring forest ecosystems, especially due to atmospheric effects.
method Estimating four vegetation indices (LAI, FAPAR, EVI, NDVI) using multitemporal Sentinel-1 SAR and ancillary data.
result Accurate estimation of forest vegetation indices using SAR data, achieving high R2 and low MAE values. New method maps land cover using radar and optical satellite images.
problem Efficiently exploiting multiple sources of information for land cover mapping.
method Deep learning framework with attention mechanism and pretraining strategy.
result Attention mechanism and extended RNN model outperform competitors.
Deep learning detects snow avalanches in SAR images with high accuracy.
problem Accurate detection of snow avalanches in remote areas using satellite imagery.
method Fully Convolutional Neural Network trained on manually labeled Sentinel-1 radar images.
result Deep learning model achieves F1 score above 66% compared to 38% for state-of-the-art methods.
The paper uses CNNs on Sentinel-2 imagery to assess landslide risks.
problem Landslide risk assessment and prediction.
method Image augmentation, 3-D CNNs, satellite imagery.
result CNNs achieve significantly better accuracy than baseline.
The paper generates a comprehensive training dataset for land cover classification.
problem Insufficient training datasets for land cover classification.
method Public Sentinel-2 data at 10m resolution matched with accurate labels, filtered and classified by Random Forests.
result Over 80% model accuracy for various locations.
New method fuses optical and SAR data to fill LAI gaps during cloudy periods.
problem Cloudy periods mask key crop growth stages, leading to unreliable yield predictions.
method Multi-Output Gaussian Process (MOGP) regression for fusing Sentinel-1 RVI and Sentinel-2 LAI time series.
result MOGP provides improved LAI estimations even during cloudy periods, especially for long gaps.
Paper uses robust GMM to reconstruct missing data in Sentinel-2 images for crop monitoring.
problem Missing data in remote sensing images, especially from multispectral and SAR sensors.
method Robust Gaussian Mixture Models (GMM) with outlier detection using isolation forest.
result Robust GMM outperforms standard GMM in reconstructing imputed values, reducing errors.
Paper improves crop classification from low-res satellite images.
problem Accurately classify land use change without high res imagery.
method Capsule layers and distributed attention with LSTM.
result State-of-the-art accuracy on crop type classification.
BreizhCrops dataset for crop type mapping from satellite imagery.
problem Classifying field crops from satellite time series data.
method Aggregated Sentinel-2 time series data, compared seven deep neural networks.
result Demonstrated the effectiveness of deep learning models for crop type mapping.
First European crop map created using satellite data.
problem Need for detailed parcel-level crop type mapping for EU.
method Used Sentinel-1 radar observations and LUCAS in-situ data.
result 80.3% overall accuracy for 19 crop types, highest for rape and turnip rape.
AI task delegation faces incentive collapse with unbounded payments as AI accuracy rises.
problem Incentive collapse in AI-assisted task delegation schemes.
method General impossibility result and sentinel-auditing payment mechanism.
result Sentinel-auditing mechanism enforces positive human effort at finite cost, independent of AI accuracy.
Paper develops a deforestation detection system using optical and SAR data.
problem Detecting tree-loss in dense forests using satellite data.
method Combines optical and SAR data, uses KL expansion for anomaly detection, and Hidden Markov Model for classification.
result Hybrid method achieves high accuracy and robustness in sparse optical data.
Deep learning improves tree species classification accuracy in imbalanced contexts.
problem Inaccurate tree species maps for large areas, especially in imbalanced contexts.
method Used deep learning models (including convolutional and attention-based) on Sentinel-2 multispectral satellite time series data.
result Deep learning models achieve higher accuracy and F1-macro scores compared to Random Forest in imbalanced contexts.
Sourcerer uses deep learning to map land cover from limited labeled data.
problem Producing accurate land cover maps with scarce labeled data.
method Bayesian-inspired, deep learning approach with a novel regularizer.
result Sourcerer outperforms other methods, even with minimal labeled target data.
Study uses satellite and lidar data to map forest height and biomass in France.
problem Mapping forest resources and carbon in large areas.
method Machine learning approach using Sentinel-1, Sentinel-2, ALOS-2, and GEDI Lidar data.
result High-resolution maps of forest height and biomass produced with good accuracy.
Detects and maps informal settlements using satellite data.
problem Mapping informal settlements for aid distribution.
method Combining satellite data and machine learning for roofing material detection.
result Effective aid distribution through accurate settlement mapping.
Detects and maps informal settlements in developing countries using satellite imagery.
problem Mapping informal settlements for aid delivery.
method Two methods: LR Sentinel-2 imagery and VHR satellite imagery.
result Successfully mapped informal settlements with LR satellite imagery.
Deep learning predicts infrastructure quality in Africa using satellite imagery.
problem Expensive and limited monitoring of infrastructure quality in developing regions.
method Convolutional neural network trained on Landsat 8 and Sentinel 1 satellite imagery.
result AUROC scores of 0.881 for Electricity, 0.862 for Sewerage, 0.739 for Piped Water, and 0.786 for Roads.
A new data set helps estimate continental-scale population distributions.
problem Lack of comprehensive, publicly available data for population estimation.
method Comprehensive data set combining satellite imagery and open-source data.
result Provides a valuable resource for developing population estimation methods.
Paper introduces kernel methods for detecting anomalous changes in remote sensing imagery.
problem Detecting anomalous changes in remote sensing imagery.
method Nonlinear extension of Gaussian and elliptically contoured distribution algorithms using reproducing kernel Hilbert space.
result Improved detection accuracy and reduced false-alarm rates compared to linear formulations.
Developed PathInf for network inference from incomplete data.
problem Massive and non-uniformly distributed missing values in data.
method Two-stage inference model: data summarization and graph inference.
result Consistently superior performance compared to state-of-the-art methods.
Study forecasts sub-city real estate prices weekly using radar and news sentiment.
problem Limited availability of reliable real estate price indicators at neighborhood and long horizons.
method Combining satellite radar signals and news sentiment to forecast sub-city real estate prices.
result The multimodal model reduces mean absolute error by 35% at long horizons (26-34 weeks).
Bayesian framework improves ML classification models' uncertainty estimates.
problem Ensuring trustworthy AI predictions with explicit uncertainty quantification.
method Proposes a Bayesian framework for generative ML classification models that accounts for input measurement uncertainty.
result The BQDA model outperforms other models in terms of interpretability, explicit uncertainty modeling, and computational efficiency.
Self-attention improves satellite time series classification without preprocessing.
problem Efficiently classifying raw satellite time series data.
method Comparison of deep learning models including self-attention, 1D-convolutions, recurrence, and random forest.
result Self-attention and recurrent neural networks outperform convolutional neural networks on raw satellite time series.
New methods detect roads in low-res satellite data, overcoming visibility challenges.
problem Detecting roads in low-resolution satellite imagery, especially those hard to see.
method Two deep learning frameworks for ordinal classification of road types from satellite time series data.
result Models can identify large and medium-sized roads from Sentinel-2 imagery.
Researchers predict butt rot volume using harvester data and remote sensing.
problem Predicting butt rot volume in Norway spruce stands for optimal forest management.
method Used random forest models with harvester information, remote sensing, and environmental data.
result Remotely sensed predictor variables were more important than environmental variables.
VDTW improves cross-year crop mapping accuracy.
problem Cross-year crop mapping accuracy is poor with existing methods.
method Vector Dynamic Time Warping (VDTW) for multi-year classification.
result VDTW achieves 99.85% and 99.74% overall accuracies for same and cross years, respectively.