D2Mamba: A mamba-based method for floodway obstructions segmentation from multispectral satellite imagery

In precision agriculture, the assessment and estimation of key crop parameters are crucial aspects for the optimisation of input usage and, as an ultimate goal, for the improvement of yield quality and quantity. In this context, a reliable prediction of yield by remotely sensed imagery is an enabling technology for optimisation. In this work, an innovative method for estimating yield in maize cultivation is presented, which exploits multi-temporal and multispectral Sentinel-2 satellite imagery with supervised Machine Learning (ML) techniques. For model training and validation, yield ground truth experimental data from combine harvesters was used, enabling the yield estimation at sub-field scale. The investigation, which was conducted on five case study plots, involved a preliminary comparison of four ML-based algorithms, trained with raw spectral bands. An assessment of the effect of the training dataset on the yield prediction accuracy was then performed. A set of Vegetation Indices (VIs) and Two Band Indices (TBIs) was also considered for this purpose. Finally, a multi-temporal analysis was conducted, in which the temporal evolution of crop spectral data over the maize growing season was exploited using imageries acquired in different epochs. The obtained results proved that an accurate estimation of maize yield can be reached using a Gaussian process regression model, exploiting multi-temporal features directly provided by the raw spectral bands. The model showed a high accuracy in the estimation of maize yield, even when fed with data acquired during only the maize vegetative phase, thus proving its capacity as a prediction tool. • Supervised machine learning techniques are used to estimate maize yield. • Combine harvester ground truth data enables prediction at subfield scale. • Multi-temporal imageries from Sentinel-2 improve the estimation. • Gaussian Process Regression algorithms reach accuracies up to R 2 > 0.9

Monitoring seasonal grass cover dynamics in a multi-use savanna rangeland is important for sustaining the coexistence of livestock and wildlife. Moreover, population growth is driving increased livestock production, which further limits resources for both livestock and wildlife. To better understand the effect of grazing on grass cover dynamics, we developed a multi-scale remote sensing approach to study the monthly variation in grass cover in two types of conservation areas: a wildlife sanctuary and a communal livestock grazing and wildlife conservancy. The study was carried out in a semi-arid region in Kenya during an exceptionally dry year of 2022 when grazing resources were limited. The Excess Greenness color index was first used to develop a model predicting green fractional vegetation cover (fCover) of field photographs. This model was then applied to upscale fCover to the landscape level using very-high-resolution Pléiades satellite data. The resulting fCover maps were subsequently used to predict grass cover from medium-resolution Sentinel-2 multispectral satellite imagery using Random Forest machine learning. The final model showed high predictive power of grass cover in May (R 2 = 0.96, root mean square error (RMSE) = 4.95%), while predictions were less accurate yet promising for January (R 2 = 0.67, RMSE = 7.1%). The monthly grass cover maps demonstrate differences between the two conservation areas; the grazing area experienced low grass cover throughout the year, whereas grass cover in the wildlife sanctuary was more driven by rainfall. The results demonstrate the usability of digital cameras as the basis for vegetation cover models. Furthermore, this method can be used for adaptive land management to monitor within-season resources for both livestock and wildlife. • Fractional grass cover can be upscaled from field to regional levels • Training data from one month was used to predict monthly variations in grass cover • Areas with higher grazing pressure had lower grass cover throughout the year • The wildlife sanctuary experienced a more natural fluctuation in grass cover

Urban planning increasingly requires multi-source geospatial intelligence to address the complexity of modern cities. This study introduces a unified deep learning framework integrating OpenStreetMap (OSM) data with multispectral satellite imagery and demographic-environmental datasets. The framework employs task-specific architectures, including Convolutional Neural Networks (CNNs) for land-use classification, U-Net segmentation for building footprint extraction, Long Short-Term Memory (LSTM) networks for traffic flow prediction, and a hybrid CNN-RNN model for air-quality forecasting. The land-use model achieved 91.6% accuracy, the U-Net building footprint extractor reached 94.0% accuracy, the LSTM traffic model obtained an RMSE of 3.6 vehicles per hour, and the hybrid air-quality model achieved an RMSE of 2.3µg/m³. These metrics reflect performance on OSM-aligned datasets specifically curated for Krasnodar. The novelty of this work lies in establishing a multi-task, multi-source analytical workflow that unifies OSM, satellite, and temporal environmental data within a coherent deep learning system. The resulting framework provides urban planners with a scalable and low-cost approach for generating high-resolution urban intelligence to support sustainable city management.

An accurate assessment of how spectral and spatial resolution influence coastal mapping remains a critical challenge for shallow-water monitoring. This study evaluates and compares hyperspectral (97 bands), multispectral (8 bands), and RGB (3 bands) data, with different spatial resolutions (10cm and 2m), to determine the most suitable spectral-spatial configuration for shallow-water mapping. For this, the effects of spectral and spatial resolution are isolated by simulating 8-band multispectral and 3-band RGB configurations at 10cm and 2m from a single 97-band hyperspectral drone dataset. This allows for comparisons between different resolutions without considering changes in temporal, atmospheric, water column, or image capture, among others acquisition-related factors. A comprehensive methodology for processing was developed using empirical and machine learning models for bathymetry estimation (Sigmoid, Subspace-KNN) and benthic mapping (SVM, FNN). The developed framework was applied at an urban sandy beach sheltered by a natural reef with rich marine biodiversity (Las Canteras beach, Gran Canaria, Spain). Results show that hyperspectral data achieved the highest accuracy (MAE, 0.15m; accuracy, 94%), while multispectral data offered an excellent balance between resolution and performance (MAE, 0.16m; accuracy, 93%). RGB data was acceptable for bathymetry but unreliable for benthic classification in complex habitats (MAE, 0.24m; accuracy, 83%). Subspace-KNN outperformed empirical models for bathymetry, and FNN improved substrate discrimination. In addition, a comparative analysis between 2016 and 2023 imagery, comparing real WorldView-2 imagery (2016; 2m and 8 bands), and drone imagery with the same resolutions emulated (2023; 2m and 8 bands), suggests an approximate 7,200m² reduction in marine vegetation that may be influenced with anthropogenic pressures and thermal increase. This approach provides a reproducible and adaptable tool for sustainable coastal management.

Abstract European coniferous forests are increasingly vulnerable to forest disturbances due to climate change and more frequent extreme events, which can trigger forest pest outbreaks. In the Italian South‐Eastern Alps, the 2018 Vaia storm led to massive outbreaks of the European spruce bark beetle ( Ips typographus L.) (Coleoptera: Curculionidae, Scolytinae), severely affecting Norway spruce forests. Sanitation felling (i.e., removal of infested trees) is a common management practice, but its efficacy in reducing bark beetle populations is variable and context‐dependent. This study assesses the effect of sanitation felling performed in 2022 in Friuli Venezia Giulia (NE Italy) on bark beetle damage that occurred in 2023 across eight spruce forest types, as evaluated using multispectral satellite imagery. Information about forest composition was retrieved from local databases to assess how forest characteristics influence sanitation felling effectiveness. Results show that bark beetle damage was effectively reduced only at very high or very low sanitation felling rates. Sanitation felling increasing up to ~50% produced an increasing damage, while damage decreased progressively only over 50% sanitation felling. Pure spruce forests, especially mountain spruce forest and spruce plantations, were the most affected forest categories, while damage was reduced in alpine (i.e., >1500 m elevation) and mixed formations. Sanitation felling above 50% proved effective in significantly reducing damage within the pure spruce neo‐reforestations and the pure subalpine spruce forests. Findings call for an integrated approach to compensate for the limitations of sanitation felling, to plan targeted interventions and help promote more resilient forest ecosystems to mitigate future outbreaks.

ABSTRACT The Thal and Cholistan Deserts in Pakistan are ecologically significant yet face numerous environmental challenges, including desertification, vegetation decline, and water scarcity. This study aims to assess the spatial-temporal dynamics and environmental challenges of these deserts through a multi-disciplinary approach. A combination of remote sensing, field surveys, and socio-economic assessments was employed. Multispectral satellite imagery from 2015 to 2023 was analyzed to detect land-use and land-cover changes, while vegetation indices such as NDVI and EVI were used to monitor vegetation dynamics. Soil and water samples were analyzed for degradation patterns, and household surveys assessed socio-economic impacts. The study revealed a 43% increase in built-up areas and a corresponding decline of 8% in vegetation cover in the Cholistan Desert, with similar trends in Thal. Groundwater levels declined by 1.5 meters per decade, and soil erosion rates were highest in areas with sparse vegetation. Socio-economic surveys showed that 70% of Cholistan households and 65% of Thal households rely on livestock and agriculture, respectively, highlighting community vulnerability to environmental changes. The findings underscore the critical impact of anthropogenic activities, overgrazing, and climate variability on these ecosystems. Sustainable management strategies integrating scientific and traditional knowledge are essential to mitigate these challenges and ensure ecosystem resilience.

Abstract Surveying an area for new fossil sites is a labor-intensive and resource-draining activity that can be alleviated with the aid of machine learning models. In karst landscapes of southern Africa, Plio-Pleistocene fossils that inform the paleoanthropological record are primarily found preserved in caves and sinkholes. The purpose of this study is to assess the utility of Random Forest (RF) models for cave and sinkhole prediction in the Cradle of Humankind, South Africa. Multispectral satellite imagery, digital elevation models (DEMs), and geologic maps were converted into raster (pixelated matrix) images in a GIS environment to denote varying aspects of the local topography, including elevation, slope, aspect, curvature, drainage, spectral reflectance, vegetation cover, fault proximity, and underlying geology. The rasters were stacked and overlaid with 1080 known cave and sinkhole locality points and 1080 random non-cave points in the study area for model training. Variable values associated with these geopoints were input into an RF model in Python for training and evaluation using a spatial ten-fold cross-validation. The model performed with 81.6% accuracy and an area under the curve (AUC) of 0.912. The importance of each variable for prediction was evaluated by measuring the increase in prediction error when variable values were shuffled. Distance to major faults, location within the Chuniespoort geologic group, dolomite presence, chert presence, and elevation exhibited the highest importance for model accuracy, while three out of 48 total predictor variables exhibited less importance than a randomly generated variable. The identification of important/unimportant variables will help build more efficient, robust models in future iterations, as well as help identify variables that could be useful in other karst regions.

Accurate estimation of forest aboveground carbon (AGC) is crucial for understanding the carbon cycle and formulating climate policies, yet it remains challenging in complex mountainous regions. This study used machine learning framework to estimate the spatiotemporal dynamics of AGC in the Three Parallel Rivers region of China from 2003 to 2024. By integrating China’s National Forest Continuous Inventory (NFCI) data with multispectral satellite imagery, we employed a two-stage feature selection strategy to identify key predictor variables. Among three ensemble algorithms tested, the Random Forest model achieved the optimal performance (R2 = 0.74). The results indicated a net increase of 67.05 Tg in total AGC over the two decades, with a spatial pattern characterized by higher densities in the west and north. Geographical Detector analysis revealed that the driving forces were synergistic, with the interaction between temperature and population density exhibiting the most prominent explanatory capacity. This study provides a high-resolution (30 m) benchmark for AGC in a global biodiversity hotspot and underscores the critical role of ecological protection policies in enhancing carbon sequestration, offering valuable insights for managing similar mountain ecosystems worldwide.

Beach morphology is influenced by climate-related changes, such as rising sea levels, shifting weather patterns, and storms, as well as human activities, making continuous monitoring essential for understanding its evolution. Within this dynamic context, some beaches develop morphological features that help attenuate the impact of high-energy events, effectively acting as natural barriers against coastal erosion and flooding. This research explores the role of Posidonia oceanica banquettes, natural seagrass accumulations, in influencing beach dynamics, shoreline stability, and dune development, processes that are common along much of the Mediterranean coast. The study developed a new methodological approach by integrating aerial ortophotos with high-temporal-resolution multispectral satellite imagery, to analyse beach evolution in the presence of Posidonia banquettes, with a focus on the impact of storm events. This approach examines shoreline, dune, and Posidonia accumulations through a combination of remote sensing techniques, enabling both medium-term through Satellite-Derived Shoreline (SDS) (∼10 years) and long-term analyses (∼70 years) using orthophotos. The results highlight the complex interactions between human activities, storm events, and natural processes, particularly the role of Posidonia accumulation in shaping beach and dune morphology. Medium-term analysis has offered detailed perspective on recent beach changes, illustrating fluctuations in Posidonia berms related to storm events and correlating shoreline positions with dune evolution. Meanwhile, long-term orthophotos analysis has provided insights into sediment transport dynamics and revealed trend patterns over extended timeframes. This integration of SDS data and aerial imagery leveraged the identification of “hotspot areas” by analysing the relationship between shoreline changes and dune toe retreat. • Shoreline monitoring using satellite and aerial images identified erosion and accretion areas. • The SDS allowed observation of shoreline fluctuations, including banquette formation. • Posidonia accumulation contributes to dune growth and erosion protection. • Kernel density maps identified dune erosion hotspots driven by shoreline interactions.

Climate change and growing populations are major challenges for food security. Understanding single-season rice (SSR) growth patterns and how much area changes over time is essential for sustaining rice distribution patterns and ensuring food security. This study utilized ground trothing data with the remote sensing (RS) technique for estimation of the SSR pattern in Jiangsu Province. A total of 1700 rice and 470 non-rice points were collected during the field visit in April-September 2023 across Jiangsu Province. The current study employed advanced machine learning (ML) and the random forest (RF) model using Google Earth Engine (GEE). This study evaluates the SSR cropping area, including the Normalized Difference Vegetation Index (NDVI), land surface temperature (LST), and land use-land cover (LULC) variation from 2018 to 2023 with different satellites. The results of NDVI show an increasing trend with mean values rising from 0.30 in 2018 to 0.42 in 2023. Additionally, higher mean values of LST were noticed in 2020 by 14.4 °C and in 2022 by 14.1 °C. Furthermore, the SSR area has significantly changed, mostly in the eastern and southern regions of Jiangsu Province, from 2018 to 2023. The higher rice cropping area decreased by 1.42% in 2019 compared to 2018, while the total reduction over the 2018-2023 period was 0.92%. Total cultivated crop areas continued to decline because most of the crop areas changed into built-up areas, resulting in a total variation of 2.75% from 2020 to 2023. The overall accuracy of RF model range was 77.33% to 93.55% with a Kappa coefficient of 0.55 and 0.87, indicating moderate to substantial classification agreement across the study period. The results of LULC indicate that the crop area decreased by 4.13% from 2018 to 2023, and major areas were converted into water bodies and built areas. In conclusion, the single-season cropping pattern decreased during the study period, accompanied by a reduction in total cropland area in Jiangsu Province. Therefore, these findings highlight the influence of urbanization and climate change on agricultural land and emphasize adaptive strategies in Jiangsu Province to ensure food security in the face of environmental challenges.

This study presents a comparative analysis of several machine learning (ML) algorithms for lithological mapping of the tectonically complex geology of the Sivas Basin, Türkiye, where salt tectonics and halokinetic deformation make spectral-based separations particularly challenging. Five target classes, including gypsum, limestone, the Karacaören Formation (Tkk), the Karayün Formation (Tsz), and vegetation were distinguished from a dataset of 1,744 endmember samples derived from Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) VNIR and SWIR bands. The five ML algorithms examined were the J48 decision tree (WEKA), Fine Tree, Support Vector Machine (SVM), k-Nearest Neighbor (KNN), and a Wide Neural Network (WideNN). The results indicate the strong potential of advanced classifiers for lithological prediction, with both WideNN and KNN achieving similar performance, exhibiting overall accuracies of 0.997. SVM also demonstrated strong generalization capability, even for high-dimensional spectral data. Feature importance analysis showed that the most influential bands were VNIR bands (B1 and B2) and SWIR bands (B4, B5 and B9), while the superior performance of WideNN and SVM was attributed to their ability to exploit additional spectral information from B7 and B8. These capabilities enabled a clear distinction between spectrally similar units such as Tkk and Tsz. Field validation based on 300 checkpoints confirmed the accuracy of the ML-based predictions, with SVM and KNN yielding the highest agreement with field observations. Overall, the results highlight the strong potential of integrating ML techniques with multispectral satellite imagery to improve lithological mapping in salt-tectonic basins and other geologically complex terrains.

Satellite-derived bathymetry holds significant value for acquiring nearshore bathymetric data. However, in coastal waters, bathymetry is affected by in-water particle scattering and seafloor substrate variability, leading to spatial inconsistency between the logarithmic green band profile derived from multispectral satellite imagery and the actual water depth profile. According to the position information of interpolated points and the inverse distance square relationship with the surrounding 16 points from low-reference bathymetric data (such as the bathymetric map from GEBCO, NOAA Electronic Navigational Charts), this model adopts a third-order inverse distance square bicubic convolution interpolation method to resample a high-resolution bathymetric map with the size of the satellite image. Normalized underwater topography trend data (derived from the low-resolution reference bathymetric map) were combined with normalized green band data to compute an averaged dataset. In this way, a linear bathymetric model was constructed. We invert this model’s parameters and calculate the water depth by using the average data and reference points from reference bathymetric data. Validation tests were conducted across three test areas using independent validation bathymetric data: Weizhou Island, China (Case II waters); Saipan, Northern Mariana Islands, USA (Case I waters); and Molokai Island, Hawaii, USA (Case I waters). Each test area was studied using five error analysis methods (i.e., scatterplot, error histogram, regional bathymetric error, three check lines, and seven check points). Compared to four classic bathymetric models (i.e., single-band model, log-ratio model, ratio-log model, and multi-band model), the proposed model achieved lower root mean square errors (RMSE) of 2.08 m, 1.40 m, and 2.01 m in the three test areas, representing reductions of 35%, 43%, 45%, and 20% and overall averages of 48%, 62%, 64%, and 43%, respectively. Its goodness of fit (R2) reached 0.87, 0.97, and 0.97, showing improvements of at least 5%, 5%, 9%, and 9% and overall averages of 17%, 77%, 84%, and 12%, respectively. The results demonstrate that the proposed model significantly improves bathymetry accuracy while maintaining algorithmic simplicity, providing a new model for acquiring nearshore foundational bathymetric maps.

Urban street trees provide essential environmental and social benefits, yet their vitality is often challenged by adverse urban conditions such as traffic emissions, impervious surfaces, and limited soil moisture. Conventional street tree management relies heavily on labor-intensive field inspections, making large-scale and repeatable assessment difficult. To address this limitation, this study proposes a GeoAI-based framework that integrates high-resolution aerial imagery, multispectral satellite data, and deep learning–based semantic segmentation to automatically delineate individual street trees and derive a remote sensing-based vitality proxy. Street trees are detected from orthorectified aerial imagery using semantic segmentation models, and vegetation indices—including NDVI, NDRE, and NDMI—are extracted from multispectral satellite imagery. An area-weighted object–pixel matching strategy is applied to associate spectral indicators with individual crowns across multi-resolution datasets. A composite vitality proxy is then constructed by integrating chlorophyll- and moisture-related indices. The results reveal spatial variability in spectral vitality signals across different urban environments. Trees along major road corridors tended to exhibit lower chlorophyll- and moisture-related indicators, while those near parks, riverfront walkways, and recently developed residential areas generally showed higher values. NDMI provided complementary insights into moisture-related stress that were not fully reflected by chlorophyll-based indices. Although the proposed vitality proxy is not a substitute for field-based diagnosis, the overall framework offers a scalable approach for citywide screening and prioritization of potentially stressed trees, supporting data-informed urban green infrastructure management within smart-city planning contexts.

Natural hydrogen has gained attention as a low-carbon energy vector, and some reported surface expressions have been linked to subcircular patterns, or fairy circles (FC), that may be detectable in multispectral satellite imagery. The Carolina Bays region, on the eastern coast of the United States, was selected because it hosts abundant, well-mapped subcircular depressions. This study aims to comparatively evaluate machine learning algorithms for identifying subcircular structures using Landsat-8 data. We processed 105 Collection 2 Level 2 scenes, masking clouds and shadows using the Level 2 quality band. Pixel-level labels were determined using a well-curated public dataset, derived from a high-resolution LiDAR survey. Traditional models (logistic regression, random forest, and multilayer perceptron) achieved precision scores below 0.66 and enabled a variable-importance analysis, which identified Band 3 (green), Band 6 (SWIR1), and five Normalised Unit Indices as the most predictive features. Deep learning models improved detection, and a U-Net architecture allowed for pixel-level segmentation of FC-like structures, producing false positives mostly in cloudy or shadowed areas. Overall, the results suggest that FC detection from multispectral data alone remains challenging due to class overlap and cloud/shadow contamination. Future work could explore integrating additional non-spectral descriptors, such as morphometric variables, to reduce ambiguities.

Nutrients play a fundamental role in crop development, and precise fertilizer application is essential for sustainable paddy cultivation and efficient resource utilization.However, traditional fertilizer management methods frequently neglect regional heterogeneity in nutrient demand, leading to inefficient nutrient use and increased environmental impact.To tackle these drawbacks, this study introduces a novel model combining Improved Multi-Scale Synchronous Convolutional Neural Network (IMSS-CNN) with YOLOv8, for intelligent nutrient monitoring and fertilizer recommendation in rice fields.The system utilizes satellite-acquired multispectral imagery to extract key vegetation indices, including NDVI (Normalized Difference Vegetation Index), GNDVI (Green NDVI), RVI (Ratio Vegetation Index), and GRVI (Green Ratio Vegetation Index), which are converted into pseudo-RGB composites for analysis.YOLOv8 is employed to detect paddy crop growth stages with high accuracy, enabling stage-specific nutrient assessment.The IMSS-CNN architecture captures deep spatial features to analyze crop health and stress, while the Botox Optimization Algorithm (BOA) further refines model performance and prediction reliability.Based on the predicted growth stage, the system calculates optimal nitrogen, phosphorus, and potassium (NPK) requirements for each field zone, providing targeted and timely fertilizer recommendations.The proposed model achieved an accuracy of 95.9%, outperforming existing approaches in both classification and nutrient estimation.This research offers a comprehensive solution to address the gaps in conventional practices by integrating satellite remote sensing, deep learning, and optimization techniques.The outcomes contribute significantly to precision agriculture by reducing fertilizer waste, enhancing nutrient uptake, increasing crop yield, and minimizing ecological harm.

Intensifying urbanisation in the Arctic, particularly in spatially constrained coastal and island cities, requires reliable information on long-term land-use/land-cover (LULC) change to assess environmental impacts and support urban planning. However, multi-decadal, high-resolution LULC datasets for Arctic cities remain limited. In this study, we quantify LULC change on Tromsøya (Tromsø, Norway) from 1984 to 2024 using a Random Forest classifier applied to multispectral satellite imagery from Landsat and PlanetScope, complemented by LiDAR-derived canopy height models (CHM) and building footprints. We mapped LULC change trajectories and examined how these shifts relate to district-level population redistribution using gridded population data. The integration of a LiDAR-derived CHM was found to substantially improve the accuracy of Landsat-based LULC mapping and to represent the dominant source of classification gains, particularly for spectrally similar urban classes such as residential areas, roads, and other paved surfaces. Landsat augmented with CHM was shown to achieve practical equivalence to PlanetScope when the latter was modelled using spectral features only, supporting the feasibility of scalable and cost-effective long-term monitoring of urbanisation in Arctic cities. Based on the best-performing Landsat configuration, the proportions of artificial and green surfaces were estimated, indicating that approximately 20% of green areas were transformed into artificial classes. Spatially, population growth was concentrated in a small number of districts and broadly coincided with hotspots of green-to-artificial conversion The workflow provides a reproducible basis for long-term, district-scale LULC monitoring in small Arctic cities where data constraints limit the consistent use of high-resolution image.

The mineral-mapping capability of three spaceborne sensors with different spatial and spectral resolutions, the Environmental Mapping and Analysis Program (EnMap), Sentinel-2, and World View-3 (WV3), is assessed regarding bauxite mining wastes in Amphissa, Greece, with validation based on ground samples. We applied the well-established Linear Spectral Unmixing (LSU) and Spectral Angle Mapping (SAM) classification techniques utilizing endmembers of two established spectral libraries and incorporated ground data through geochemical and mineralogical analyses, X-ray fluorescence (XRF), Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS), and X-ray Diffraction (XRD), to assess classification performance. The main lithologies in this area are bauxites and limestones; therefore, aluminum oxyhydroxides, calcite, and iron oxide minerals were the dominant phases as indicated by the XRF/XRD results. Almost all target minerals were mapped with the three sensors and both methods. The performance of EnMap is affected by its coarser spatial resolution despite its higher spectral resolution using these methods. Sentinel-2 is most effective for mapping iron-bearing minerals, particularly hematite, due to its higher spatial resolution and the presence of diagnostic iron oxide absorption features in the VNIR. World View 3 Shortwave Infrared (WV3-SWIR) performs better when mapping calcite, benefiting from its eight SWIR spectral bands and very high spatial resolution (3.7 m). Hematite and calcite yield the highest accuracy, especially with SAM, indicating 0.80 for Sentinel-2 (10 m) for hematite and 0.87 for WV3-SWIR (3.7 m) for calcite. AlOOH shows higher accuracy with SAM, ranging from 0.57 to 0.80 across the sensors, while LSU shows lower accuracy, ranging from 0.20 to 0.73 across the sensors. This study showcases each sensor’s ability to map minerals while also demonstrating that spectral coverage and the spatial and spectral resolution, as well as the characteristics of the selected endmembers, exert a critical influence on the accuracy of mineral mapping in mine waste.

This article deals with a calling issue of sustainable development in regions with mining industries. The authors developed an algorithm and presented the architecture of a software system for automated remote monitoring of man-made land problems. It is based on comprehensive application of modern machine training methods, specifically self-supervised learning (Dino, MAE, MoCo) and the Vision Transformer architecture for the analysis of multispectral satellite imagery from open sources (Sentinel, Landsat). The proposed solution automates the entire data management cycle from gathering and preprocessing to segmentation of objects (quarries, waste heaps), area measurement, and time-series analysis of spectral indices, such as the normalized difference vegetation (NDVI) and the normalized difference water (NDWI) ones, which serve as indicators of vegetation cover and water body condition, respectively. The system is integrated with GIS via a QGIS module, and its functionality is accessible via an API, ensuring easy use and integration into existing workflows. The proposed approach provides increased efficiency and accuracy of monitoring, making a basis for informed management decisions in the area of rational employing natural- and subsoil resources

ABSTRACT Fusing hyperspectral (HS) and multispectral (MS) satellite imagery enables the joint exploitation of fine spatial detail and rich spectral information, supporting accurate environmental and geological analyses. Deep Learning methods have recently become the top-performing solutions for HS–MS fusion, yet their progress is limited by the scarcity of large, diverse training datasets. Existing datasets, mainly derived from airborne acquisitions, cover small regions and lead to poor generalisation and network overfitting. To address this, we introduce a physics-based framework that generates realistic HS–MS training data directly from satellite sensors. Applied to Sentinel-2 (S2) and EnMAP, it combines S2’s spatial detail with EnMAP’s spectral signatures while simulating each instrument’s Point Spread Function, Modulation Transfer Function, and Spectral Response Function. This approach enables the creation of large, physically consistent, globally distributed datasets with strong generalisation capability. A transformer-based autoencoder trained on these synthetic data achieves high spatial and spectral fidelity in S2–EnMAP fusion. Comparable performance obtained with other Deep Learning architectures confirms that the physics-grounded supervision, rather than the network design, is the main driver of success. Validation on real Sentinel-2 and EnMAP imagery demonstrates the robustness of this scalable, sensor-specific framework for practical, global HS–MS fusion.