vegetation index calculation

This study addresses the fragmented and irreproducible nature of existing hyperspectral data processing pipelines in agriculture, which lack unified open-source tools to support plant phenotyping research. To bridge this gap, the authors present the first end-to-end, open-source Python library for leaf-level hyperspectral preprocessing. The library integrates raw ENVI data calibration, leaf detection and cropping based on vegetation indices such as NDVI, CIRedEdge, and GCI, data augmentation, and spectral visualization capabilities. It supports both command-line execution and programmatic import as a library module. By standardizing the preprocessing workflow, this work significantly enhances consistency and efficiency in hyperspectral data handling for plant phenotyping, enabling reproducible and scalable analyses.

To address the challenge of balancing interpretability and discriminative power in spectral indices for vegetation remote sensing classification, this paper proposes a novel method for automatically constructing concise, illumination-invariant polynomial spectral indices. Starting from the normalized difference as a fundamental building block, we perform polynomial expansion to generate physically meaningful candidate features. We then integrate ANOVA-based filtering, recursive feature elimination, and L1-regularized SVM to achieve rigorous, statistically driven sparse feature selection—marking the first deep integration of polynomial index construction with data-driven feature selection. The resulting minimal indices (e.g., b5×b6 using Sentinel-2 red-edge bands) are derived exclusively from interactions among Sentinel-2 bands b4–b8. In the Kochia identification task, a single optimized index achieves 96.26% accuracy, while an ensemble of eight indices attains 97.70%. All indices are lightweight and fully compatible with real-time deployment on Google Earth Engine.

To address the latency of traditional ground-based monitoring in rapidly evolving wildfire scenarios, this study proposes a synergistic remote sensing analytical framework integrating multispectral aerial and satellite imagery. Methodologically, we systematically evaluate and optimize the joint applicability of the Normalized Vegetation Difference Index (NVDI), Modified Normalized Difference Water Index (MNDWI), and Soil-Adjusted Vegetation Index (MSR) for segmenting critical wildfire environmental elements—flammable vegetation, water bodies, and built-up structures—and combine these indices with spatial analysis to achieve high-accuracy land-cover classification. Validation on two real-world wildfire events demonstrates that the framework improves fire-hazard element extraction accuracy to over 92% and reduces early-warning latency by 40%. These advances significantly enhance both the precision and operational feasibility of dynamic fire-risk assessment and response decision-making.

This study evaluates the capability of DESIS spaceborne hyperspectral data to predict plant species richness across two representative habitats in southeastern Australia—Southern Tablelands and Snowy Mountains. We systematically compare three feature extraction methods (PCA, CCA, PLS) coupled with three regression models (KRR, GPR, RFR), employing two-fold cross-validation. Results indicate that spectral bands in the red-edge, red, and blue regions contribute most substantially to prediction performance. DESIS-derived estimates significantly outperform those from Sentinel-2: in Southern Tablelands, r = 0.76 and RMSE = 5.89; in Snowy Mountains, r = 0.68 and RMSE = 5.95. This work demonstrates the feasibility and superiority of spaceborne hyperspectral data for regional-scale remote sensing of plant diversity, providing both methodological guidance and empirical validation for biodiversity assessment using next-generation hyperspectral satellites.

This study addresses the challenges of sparse and irregular satellite NDVI observations caused by cloud cover and the difficulty of short-term forecasting of crop vegetation dynamics under heterogeneous climatic conditions. The authors propose a probabilistic forecasting framework that employs a deep learning architecture to separately encode historical NDVI and meteorological observations along with future exogenous covariates, fusing multimodal information for multi-step quantile prediction. A novel temporally distance-weighted quantile loss function is introduced, complemented by feature engineering that incorporates both cumulative and extreme weather metrics, effectively capturing the delayed vegetation response to meteorological drivers and temporal uncertainty. Experiments on European satellite data demonstrate that the proposed method outperforms existing statistical, deep learning, and time series baselines in both point and probabilistic forecasting metrics, with ablation studies confirming historical NDVI as the dominant predictor and meteorological covariates providing significant performance gains.

This study addresses a critical gap in remote sensing research—the lack of large-scale, physically consistent synthetic datasets that combine high spectral resolution with pixel-level ground truth for vegetation traits. To this end, we integrate the PROSAIL radiative transfer model with Sentinel-2 Level-2A inversion results to generate, for the first time, physically realistic hyperspectral image cubes (400–2500 nm, 64×64 pixels) across four ecologically distinct regions, yielding 10,915 samples. Each sample includes paired pixel-level vegetation trait maps, uncertainty bounds, and scene classification layers. The resulting dataset enables rapid radiative transfer simulations, benchmarking of inversion algorithms, and investigations into spectral–biophysical relationships, offering a high-quality, verifiable reference resource for advancing remote sensing modeling and machine learning applications.

Existing vegetation prediction models are constrained by fixed observational meteorological trajectories, limiting their capacity for multi-scenario response analysis. This work proposes the first geospatial world model, which integrates sparse NDVI time series, historical meteorological covariates, and static geographic context through a recurrent latent dynamics architecture. Without requiring scenario-specific supervision, the model simultaneously enables probabilistic forecasting under observed conditions and conditional simulation under user-defined meteorological forcings. Evaluated on the GreenEarthNet benchmark, it significantly outperforms both temporal and remote sensing baselines, achieving state-of-the-art performance in both point and probabilistic prediction tasks. The model successfully reproduces vegetation dynamics across four future climate scenarios in Europe, with summer 2022 simulations over France aligning closely with known temperature–moisture sensitivities, thereby demonstrating strong spatiotemporal generalization capabilities.

This study addresses the challenges of low computational efficiency and poor model scalability in pixel-level cross-temporal classification for high-resolution vegetation monitoring. To this end, the authors systematically optimize the Vision Transformer architecture and, for the first time, demonstrate its efficiency and scalability in vegetation phenology monitoring. Through comprehensive design explorations across seven dimensions—including data normalization, spectral ordering, boundary handling, spatial windowing, tokenization strategy, positional encoding, and feature aggregation—and supported by multidimensional ablation studies, the proposed method is validated on UAV and ground-based imagery from the Brazilian Cerrado biome. Compared to multi-temporal CNN baselines, the approach reduces FLOPs by an order of magnitude while maintaining competitive classification performance, and its parameter count remains constant regardless of sequence length, making it well-suited for resource-constrained phenological monitoring scenarios.

Effective forest health monitoring requires high-resolution, large-scale remote sensing approaches to address both natural and anthropogenic disturbances. This study leverages 10-meter Sentinel-2 satellite data, integrating ecological and topographic contextual information with vegetation phenological cycles to develop a conditional quantile regression model based on the Normalized Difference Vegetation Index (NDVI). For the first time, this approach enables national-scale (Switzerland) detection of forest greenness anomalies at 10-meter resolution. The method substantially improves modeling accuracy during the greening-up period, produces spatially coherent browning anomaly maps, explains 65% of the variability in the median seasonal cycle, and successfully identifies multiple known disturbance events. These results provide quantitative assessment and visual support for forest conservation efforts.

This study addresses key challenges in fine-grained vegetation community classification—namely, overlapping class characteristics, insufficient probability calibration, weak evaluation of minority classes, and poor model stability—by proposing the Calibrated EcoTreeFuseNet-Plus framework. This approach fuses output probabilities from tree-based models and neural networks, integrating out-of-fold predictions, meta-learning, and temperature scaling to effectively mitigate stacking leakage while maintaining strong discriminative performance and well-calibrated probabilities even with limited samples. The method leverages LiDAR-derived terrain and canopy variables alongside hyperspectral vegetation indices, achieving a test-set accuracy of 0.8000, a macro F1-score of 0.7768, and a remarkably low expected calibration error of 0.0651. Consistency across five random trials is evidenced by a macro F1 standard deviation of only 0.0112, demonstrating substantial improvements in accuracy, calibration, and stability.