Earth Resources and Environmental Remote Sensing/GIS Applications XV

An operational product dedicated to the Portuguese forest sector in under development. The goal is to provide up-to-dated, open and friendly spatial information representing vegetation loss across forest and shurblands. Sentinel-2 data were used to detect vegetation loss every two months over a period of two years. Expeditious expert-knowledge rules were applied to NDVI time series to detect the location and date of potential changes, without the need for complex or time-consuming methods. Preliminary results reveal omission and commission errors of approximately 19% and 15% and small computational requirements. The final product is a user-friendly layer in vector format available through an online viewer and WMS service of the Portuguese Land Cover Monitoring System SMOS, updated bimonthly and early users include private companies and public institutions.

Forest disturbance dynamics are changing in forests across continents in response to global change. Detailed quantitative data on past disturbance events are necessary to establish baselines against which to assess change. Here, we used dense Landsat time series and a novel forest monitoring algorithm (Continuous Change Detection and Classification - Spectral Mixture Analysis) to monitor forest disturbances in Spain from 1985 to 2023 and adapted the algorithm separately for two biomes (temperate and, for the first time, Mediterranean). We characterised disturbance regimes at the national scale and by the two biomes and forest types (needleleaf, broadleaf and mixed). We found that the total extent of disturbed forest had previously been underestimated by existing products and accounted for 4.5 million ha, and that disturbance size and severity were significantly different (i.e. using the Wilcoxon signed-rank test, p < 0.05) between forest types and biomes, but frequency only for biomes.

This paper addresses the challenges of visually interpreting PolInSAR data for environmental monitoring. PolInSAR data provides valuable information on geophysical properties such as soil moisture, vegetation height and density. While automated techniques can be used for land cover classification, human visual interpretation remains crucial for considering contextual information and integrating domain knowledge in data exploration. However, visual interpretation of PolInSAR data is challenging as a single image representation captures only a fraction of the information. To address this, we propose an approach combining polarimetric and interferometric feature extraction and dimension reduction techniques. By projecting features into a 3D space using UMAP, we generate a comprehensive image representation facilitating land cover identification. Applied to PolInSAR data from the Wadden Sea, our approach enables easy recognition of various types of salt marshes, dune’s vegetation, and tidal flats. This supports effective monitoring and aids decision-making for conservation efforts.

Wetlands are particularly vulnerable to human activities. The ecosystem is very unique due to the comparison of wetlands to other areas. Therefore, human activities are of high importance to the biodiversity of wetlands. Quantifying the impacts of human activity on wetlands requires monitoring of both wetland extent and dynamics, on the one hand, and human activities, such as those connected to urbanisation, on the other hand. The use of different satellite Earth observation data can go a long way in quantifying the relationship between wetlands and human activities. This paper aims to use different satellite data over a long period of time, including optical data, nighttime remote sensing data, to invert changes in wetlands and changes in human activities. Furthermore, the results will contribute to a better understanding of the relationship between human activities and the environment, and a better calibration of nighttime RS data.

In this study we propose a framework to automatically monitor macroplastic loads in rivers using Deep Learning. The approach was evaluated with a measurement campaign at the Rhine river (in Koblenz, Germany). An RGB camera was installed on a bridge and captured images of the objects in the river. Various plastic and vegetation objects in different degradation states were introduced upstream and recollected downstream. Our dataset consists of about 800 images with objects labelled in three categories (plastic bottles, plastic litter, vegetation). The pre-trained YOLOv5 network showed promising validation results with a mean average precision (mAP@0.5) of about 94 %.

Reducing carbon emissions in the built environment is crucial to achieving the United Kingdom's target to become Net Zero by 2050. In this article, building physics and electric energy systems modelling techniques are used along with publicly available Ordnance Survey data to generate geographic data sets intended to be used by local authorities and distribution network operators in Local Area Energy Planning. The Local Energy Net Zero Accelerator (LENZA) project, an effort involving a range of stakeholders including local authorities and the private initiative alike, is presented as a use case of such data sets in a single tool to determine the retrofit and rollout potential of domestic heat pumps and inform the creation of a low-carbon heating strategy in the city of Dundee.

In the last few years, “NewSpace” has become the notation to describe the boom in satellite based remote sensing, caused by already established commercial players as well as start-ups, causing a high dynamic in developing and introducing innovated technologies. Focusing on Synthetic Aperture Radar (SAR) remote sensing, NewSpace stands for all missions that enable very frequent and enhanced imaging being available at affordable costs. To ensure the high imaging frequency, so-called formations or constellations of satellites have become state of the art. Here, we considered free-available NewSpace SAR amplitude imagery to apply an established change detection approach. By adopting the approach on this data, the functionality is proven and parameter are identified which have to be adjusted. The robustness of the change detection results is evaluated by a supervised comparison with optical data if available.

It is challenging to interpret line-of-sight displacement from Synthetic Aperture Radar (SAR) satellite for infrastructure monitoring. 2.5-dimensional analysis, a previous method to decompose line-of-sight, has limitations when it’s applied to infrastructure. Data must be acquired by two satellites under the specific situation. To address these issues, this research proposes a method to decompose the line-of-sight displacement into vertical and bridge-axis displacements using a single satellite, overcoming the limitations of 2.5-dimensional analysis. The method introduces a constraint that the displacement vector follows the deformation model of a bridge, and it has been validated with both simulation and measured data.

Coastal areas are very complex and dynamic environments where protection of valuable and vulnerable habitat is a priority, consequently principles of sustainable engineering must be adopted. Coastal erosion, aggressive pollution of sensible ecosystems, unstable slopes and cliffs, adverse chemical conditions related to the high concentration of chlorides, presence of cultural heritage or fragile natural sites, hard impacts of human activities and anthropization are only some examples of the severe issues that typically have to be considered and tackled. This study presents some preliminary results of integrating multi-level information for the creation of a checklist of useful indicators, suitable to define an early-warning index for the infrastructure and the areas concerned, to guide detailed on-site analysis with more conventional inspection techniques such as total stations or drone inspections. MT-InSAR was used to identify displacements over transport infrastructure. The NDWI and NDMI were used to monitor variations of the coastal area over time, while the SWIR indicator allows the detection of changes in land cover enabling infrastructure and coastal dynamics monitoring.

Bragança (Portugal) has a network of in situ sensors, which measures the Air Temperature (Tair) at 23 points classified into seven different Local Climate Zones (LCZs). The main objectives of this study were: i) calculate the LST with data from the ASTER sensor on Google Earth Engine (GEE) at these 23 points, between 2000-2023, divided by seasonality: (a) summer/spring; and (b) autumn/winter; ii) calculate the intensity of SUHI (SUHIint) and UHI (UHIint), using the LST and Tair, respectively; iii) correlate SUHIint and UHIint; and iv) evaluate the distribution of temperatures in the LCZs. Both in (a) and (b) the highest medians of SUHIint refer to classes with anthropogenic elements, and (a) presented greater heterogeneity in the results. The correlation of Tair and LST between 2011-2023 was "strong" for 74% of the points and "very strong" for 26%.

This study presents an advanced approach for live fence detection using convolutional neural networks (CNN) applied to high-resolution satellite imagery. Preprocessing was conducted with Google Earth Engine (GEE) to enhance image quality. The CNN model, trained on diverse datasets from multiple providers, achieved high precision (0.85), recall (0.91), and F1-score (0.88). Comparative analyses showcased the efficacy of our approach in accurately detecting live fences across various geographical regions, contributing to sustainable land management practices.

The Rias Baixas region, in NW Spain, is home to significant heavy-mineral placer deposits, classified as Critical Raw Materials by the European Union (EU). This study, conducted within the S34I Project framework, funded by the EU, utilized Google Earth Engine (GEE), a cloud-based platform, to perform automatic lineament extraction in the Ria de Vigo, Spain. For this purpose, a pre-processed Sentinel-1 image was retrieved from GEE’s collection. Initially, the Canny edge detection algorithm was employed, followed by the Hough Transform algorithm for line detection. Lastly, an analysis was conducted on the length and orientation of the extracted lineaments. GEE has shown to be an efficient and accurate tool to automatically extract lineaments, which are crucial for comprehending the study area’s geological context and useful for mineral exploration.

Remote sensing offers a cost-effective solution for monitoring water resources across vast areas and timeframes. This study introduces an innovative framework within Google Earth Engine (GEE) to process high-quality Sentinel-1&2 data for detecting and monitoring water surfaces. The approach focuses on neglected small water bodies in semi-arid regions of SW Iberia. By employing Sentinel-1&2-based local surface water (SLSW) models, this research surpasses existing methods in accuracy. The comparison with Landsat-based global water (LGSW) models indicates a SLSW superior performance in capturing seasonal patterns and aligning with validation data. The findings underscore the importance of understanding dynamic surface water characteristics, especially in small-sized water bodies crucial for ecological systems. This research contributes to sustainable water management strategies, aiding in identifying anomalies and supporting rural development and biodiversity conservation efforts. The proposed approach holds significant potential for addressing water scarcity challenges in regions susceptible to climate change and agricultural intensification.

Gas flares onshore are mixed with other industrial infrastructure, artificial light and buildings making detection exceptionally complex. In this study, we presented a detecting method for gas flaring onshore using the VIIRS NightFire data, the Landsat-8 OLI and Sentinel-2 MSI data. The results showed that the detecting method presented in the study has high performance. It not only could offer a reliable way to detect gas flares onshore, but also provides effective measurements to control methane emissions.

Floods are considering as one of the most common natural disasters, affecting seriously both natural and man-made environment. Remote sensing has proved to be an effective solution for post-disaster mapping and monitoring, facilitating the proper management of such crises. In this framework, the current research focuses on the post-disaster mapping of the flooded areas in the wider region of Thessaly using multispectral and radar data of Sentinel-1 and Sentinel-2 missions respectively. Sentinel data were processed using two different approaches, i.e. a manually one and a more automated one performed using GEE, while flood mapping was based on various methodologies such as simple digitization, thresholding, MNDWI calculation and random forest classification.

Aerial mapping relies on accurate external camera orientations (EO) for georeferencing and orthoprojection. After the 2021 Ahrtal flooding in Germany, a fixed-wing drone captured oblique images of the expansive area, including flooded regions. We employ machine learning for road detection on these images and OpenStreetMap (OSM) reference data to optimize EO parameters. We thereby enhance image projection accuracy onto an ortho mosaic.

New remote sensing platforms such as Unmanned Aerial Vehicles and innovative processing techniques such as computer vision, opened up new horizons and perspectives in landslide monitoring research via 3D point clouds and multiscale model-to-model cloud comparison algorithms to evaluate landslide evolution. Those algorithms detect landslide scarps, landslide evolution, displacement rates etc. In steep terrains the UAVs seems to be an efficient way to obtain accurate 3D data. This case study refers to the Panagopoula landslide, an active landslide developed in narrow and steep valley. Nadir and oblique imagery acquired from UAVs and ground control points collected with RTK GNSS were used to reconstruct the landslide.

The area of Western Greece is characterized by numerous faults, high seismicity, and ongoing land deformations. To better understand the behavior and patterns of land displacements in this region, we combined Sentinel-1 Interferometric Synthetic Aperture Radar (InSAR) data with ground-based measurements from Global Navigation Satellite System (GNSS) permanent stations. InSAR data were processed with two state-of-the-art workflows and compared with estimates from the European Ground Motion Service. The magnitude of these displacements was further scrutinized by examining the derived time series of land deformations at selected areas with critical infrastructures, shading new light on infrastructure monitoring from space.

Landslides present intricate challenges for prediction, often leading to devastating consequences. The difficulty of measuring parameters during these events impedes the characterization of slope failure mechanisms. Active remote sensing, primarily using InSAR, has proven effective for large-scale kinematic characterization and monitoring of landslides. This study aims to apply multi-temporal analysis to characterize displacement patterns of catastrophic landslides and identify susceptible hillslopes prone to rapid failure. Displacement analysis utilized PSI and SBAS methods on Sentinel-1 data. This research contributes to managing landslide risks by offering improved data to enhance mitigation strategies, formulate evacuation plans, and support recovery efforts. Furthermore, the displacement measures and patterns over government-known and unknown areas enhance early warning systems for landslides, which is crucial in developing countries to build resilient communities and guide territorial planning.

The present study focuses on exploring two decision tree machine learning algorithms, XGBoost and Random Forest (RF), to predict alteration zones associated with Cobalt-Nickel mineralization in Asturias, located in the northwest of Spain, at the Saint Patrick License area, which encompasses the Aramo plateau. The training dataset was extracted of the bands of Landsat-9 and PRISMA satellites using alteration zones that were previously identified by AURUM company. To reduce the dimensionality and increase computational efficiency, Independent Component Analysis (ICA) was applied to the satellite bands. As result, the pixels of the image were classified as host rock or alteration zone. For the PRISMA image, the RF algorithm achieved a mean classification accuracy of 0.97. The accuracy for the Landsat 9 image was at 0.90. The XGBoost algorithm demonstrated an accuracy of 0.95 for the PRISMA image and 0.82 for the Landsat 9 image, indicating reduced overfitting.

Santa Marta Beach is located in Ría de Vigo (NW Spain), in an estuary environment where placer deposits were reported. This study, conducted within the S34I Project framework, funded by the EU, employed aerial imaging study (images from the National Aerial Orthophotography Plan, and an ad hoc UAV survey) and geological field observations in coastal areas. The aim of this work is to map the inland-sea transition zone and improve the knowledge on the relationship between geological features and CRM placer deposits. To this end, we have carried out a reconnaissance of fractures, veins and placer areas around Santa Marta Beach, integrated into a Geographic Information System (GIS). Due to their high resolution, the aerial images provide a good structural and placer reconnaissance of the coastal areas.

• The Central Iberian Zone (CIZ) of the Iberian Massif is renowned for its wealth of gold deposits, with the Freixeda deposit in the Mirandela district of Northern Portugal standing out due to its economic and strategic value. However, gold-antimony (Au-Sb) exploration in the Freixeda area remains under-researched. This study focuses on using specialized Band Ratios (BR) to identify and map hydrothermally altered zones that signal the presence of Au-Sb mineralization in the region. To achieve this, we utilized hyperspectral data from the PRISMA satellite, which was recently launched and offers advanced spectral analysis capabilities. The BR methodology adopted in our research derives from targeted feature extraction indices applied to this hyperspectral data. The findings of the study indicate that the PRISMA satellite's hyperspectral imaging is a powerful asset in pinpointing Au-Sb mineralization within the Freixeda deposit. These insights contribute to more informed and targeted exploration initiatives in the future.

The present study evaluates the 3D point clouds derived from high accuracy Unmanned Aerial Vehicle’s (UAV’s) cameras to the respective data collected by Terrestrial Laser Scanner (TLS). An open pit bauxite quarry in Greece monitored in the frame of “m4mining” project is selected as the study area. “Μ4mining” is an EU-funded program aiming at confining the resolution gap between satellite and UAV-acquired data for mine monitoring. The 3D point clouds derived from UAV flight campaigns and TLS measurements were compared in terms of point density and fidelity of topographic representation.

This study examines the efficacy of Harmonized Landsat Sentinel-2 (HLS) data in copper mineral exploration, particularly through the application of remote sensing band ratios. By analyzing the spectral characteristics of HLS data, suitable bands for ratio computations are identified to enhance the detection of mineral alteration. Moreover, statistical analyses, including normal distribution statistics and Pearson correlation coefficients, are employed to assess and compare spectral enhancements for specific mineral segments. The results are validated and complemented using Principal Component Analysis (PCA). This systematic approach offers a promising method to improve the efficiency of copper mineral exploration, providing valuable insights for geological surveys and resource management.

The Kurdo-Zagrosian mountains of Marivan in western Iran have experienced numerous fires in recent years. Mapping forest fire susceptibility is crucial for prevention and policy development. Machine Learning (ML) plays a key role in remote sensing applications related to fire detection, severity assessment, and prediction. This study utilizes Non-negative Matrix Factorization (NMF) to detect fire-susceptible areas in Kurdo-Zagrosian forests of Marivan and Sarvabad in Kurdistan Province, Iran. NMF, a ML method, enforces non-negativity constraints on factor matrices, making them interpretable. Sentinel 2 satellite imagery, elevation, distance from road network, and Zagros Grass Index (ZGI) are used as inputs. Results indicate NMF efficiently handles large-scale datasets and identifies high fire-susceptible regions accurately, showing significant overlap with fired areas.

We present a new biodiversity monitoring tool in Google Earth Engine to measure trends in species habitat suitability over time using ecological niche models (ENMs) with a time series of satellite products. Focusing on Montesinho Natural Park in northeastern Portugal, the application uses MaxEnt to calculate species distribution models for amphibians, birds, mammals, vascular plants, and reptiles with data from six Moderate Resolution Imaging Spectroradiometer (MODIS) products from 2001 to 2023.

Snow is the component of the Cryosphere with the largest seasonal variation in spatial extent. In fact accumulation and rapid melt of snow are two of the most dynamical seasonal environmental changes on the Earth’s surface. The large scale changes in snow cover are useful as indicators of climatic changes, snow also affects other components of the Earth system at a variety of scales. The main aim of the presented research is to trace the use of different satellite data and approaches to track the dynamics of the development of the snow coverage.The subject of the study is snow coverage and its dynamics for different seasons around Vitisha, Rila and Pirin Mountain. The objects were analyzed and mapped according to Еuropean Space Agency data ( ESA )- Copernicus program.Results have been obtained for quantitative changes of wet snow cover and its dynamics. The data used are with a high time-spatial resolution. The SAR capabilities for snow monitoring are known to be extremely effective in terms of observation frequencies . The snow mapping system has sufficient time and spatial resolution.

As climate change is now in full swing across the globe affecting both inland and coastal zones, it is necessary a database of the impact of the various factors affecting shoreline erosion to be created. Physical factors, such as the geology of the area, the coastal slope, the average wave height, the coastline evolution, the mean tidal range, and the sea level rise (SLR) can be enumerated, through remote sensing techniques, and then be included in the Coastal Vulnerability Index (CVI) formulas providing the rate of threat a coastal zone faces. The present study examines the efficiency of six CVI formulas in coastal hazard determination of the littoral village of Arkoudi, situated in the north-west shore of the Prefecture of Ilia’s, Greece. High – resolution (HR) orthomosaics and Landsat datasets were used and compared. The results show that these formulas calculating using high – and low -resolution dataset, yield a mean declination ranging from 0.12 to 0.98 units, depending on the formula.

The coastal zone is a dynamic environment where human structures and activities take place near or along it. The reduction of sediment supplying the coast through rivers due to human activities in conjunction with the climate change is expected to trigger changes in the prevailing coastal regime, causing failures of existing infrastructure, affecting the local ecosystem and the local economy. The present study examines the coastal evolution in the littoral village of Spiantza located in the prefecture of Ilia, Greece, after the construction of the Alfeios dam in 1962. More specifically, the shoreline change rates of the 1945-1996 period were compared to the respective rates of the 1996-2016 period, utilizing remote sensing data and techniques, while the future shoreline position of the year 2036 was estimated. The results show that since 1996 the wider area has suffered severe erosion reaching –2.0 m/yr, covering almost 95% of the area, affecting the littoral zone equilibrium, and threatening coastal structures. Additionally, the forecast model indicated that the mean rate of erosion is expected to increase to –2.93 m/yr by the year 2036.

The use of active sensors like Light Detection and Ranging (LiDAR) to generate very high resolution Digital Terrain Models (DTM) which can help to detect terrain related mineral alteration zones when using a multitemporal approach. In this study three datasets were used, the earlier studies from 2012 and 2021, provided by the Spanish National Geographic Institute (IGN), and the third one, from 2023, provided by the S34I Project – Secure and Sustainable Supply of Raw Materials for EU Industry, a HORIZON Project. This study analyses terrain anomalies related to mineralization and alteration within a karst environment, potentially related to Copper-Cobalt-Nickel mineralization, in a study area located in Asturias, Northwest Spain. After processing the data several anomalous height differences were identified which require validation in the field , to check and verify the nature of these occurrences.

Abstract: Serious natural catastrophes like landslides may cause significant damage to houses, businesses, and infrastructure in addition to fatalities. There has been a noticeable increase in the frequency of landslides over the last several decades, which may be linked to both urbanization and climate change. One of the best tools for examining landslides is remote sensing. These methods are often divided into three main categories, albeit there may be some restrictions on this categorization and ambiguity in the borders. The first lesson covers techniques for locating landslides, namely the mapping of past or present slope failures. The monitoring of landslides, which includes measuring ground deformation and looking into any other temporal changes like plant cover and land usage, is the second component. Landslide analysis and prediction approaches are included in the third class. This paper offers a thorough summary of the three different landslide investigation types that use remote sensing techniques. The research concludes with a unique classification of Terrestrial Laser Scanning (TLS) and Unmanned Aerial Vehicle (UAV) techniques that are appropriate for investigating va

WACAPOU project (Water Awareness and Compensation Assessment Program for its Optimum Use) participates in reducing water footprint of activities focus on: · a measurable representative water footprint calculation · compensation solutions to replenish and/or maintain the freshwater resources. Green water is the biggest part of water, representing about 60% of rainfall. This is precipitation on land, not running off as opposed to blue water, but stored in the soil, evaporated, or transpired through plants. And it is not included into Water Footprint calculation. The Objectives of green water quantification: integrating green water into the water footprint calculation and promoting a local water cycle. The main difficulty is to measure green water in term of volume per unit of time. The aim of the work is to explore different methodologies proposed in the literature to establish a valid green water model based on remote sensing data (satellite images) and UAV (drone) acquisitions supported by field data in order to monitor different areas of interest.

he VIIRS instrument onboard the NOAA-21 spacecraft has successfully operated since its launch on November 18, 2022. A panchromatic channel in VIIRS, referred to as the day-night band (DNB), was designed with multiple gain stages resulting in a large dynamic range and high sensitivity such that its detectors can make observations during both spacecraft day and spacecraft night. The on-orbit calibration performance is monitored via the gain trending of low-gain-stage (LGS), gain ratios of mid-gain-stage (MGS) to LGS, and high-gain-stage (HGS) to MGS, as well as dark offsets in all modes, detectors and HAM sides. Contamination of DNB images due to straylight has been observed in previous VIIRS builds. Data from monthly new Moon observations have been used to estimate the straylight impact such that a look-up-table (LUT) has been built each month to correct the contaminated images. In this paper, the calibration algorithm and performance of NOAA-21 VIIRS DNB have been presented, together with the comparisons to previous VIIRS instruments onboard the SNPP and NOAA-20 spacecraft. As we strive to continuously improve the straylight correction, we will explore the possible straylight estimation improvements, including investigations on straylight dependence of solar zenith angle and solar azimuth angle by entering the instrument from both the solar diffuser and the Earth view ports.

This research investigates the effectiveness of the SHAP (Shapley Additive exPlanations) approach in enhancing the interpretability of landslide susceptibility models. Focused on the landslide-prone region of Bhutan, the study compares the performance of two settings: one incorporating a physical-based model and the other without. Through various evaluation metrics, including overall accuracy and precision-recall, the study assesses the predictive capabilities of each model. The findings illuminate the strengths and limitations of both approaches, offering valuable insights for stakeholders and decision-makers involved in land use planning and disaster preparedness. Ultimately, this research aims to advance the field of landslide susceptibility modeling by elucidating the role of SHAP and its interaction with physically based models, thereby contributing to more effective risk mitigation strategies in challenging terrains.

“XR-EVPDM” (XR for Evacuation Planning within Disaster Management), A system for evacuation planning within disaster management, This innovative platform include XR (Extended Reality), AR (Augmented Reality), VR (Virtual Reality), digital elevation models (DEM), weather attributes such as :- temperature, precipitation, wind speed disaster attributes includes type, intensity, frequency of past events and precise geographical coordinates (longitude and latitude) to revolutionize the way evacuation strategies are conceived and executed.

Susceptibility mapping using remote sensing is one of the most important tools for assessing natural hazards. These maps can provide valuable information about areas prone to certain hazards. Two methods, the Analytical Hierarchy Process and Machine Learning based models have been deployed over the years for producing susceptibility maps, with each method having its own advantages and limitations. Generating input data for the above two methods is time and effort intensive. In our current work we propose a unique technique which combines both AHP and ML-based models. Our main objective is being able to accurately detect past natural hazards for a given region, while optimizing the time and efforts required to generate input data. The results are promising, with the performance of our proposed hybrid model surpassing the performance of a AHP model while being comparable to the performance of the ML model while requiring less than half the efforts to generate input data as compared to the two models.

Airborne Laser Scanner (ALS) has become one of the most popular LiDAR systems in the last decades for studying forest structure and ecosystem dynamics. A systematic review of scientific literature was performed to assess the use of ground filtering algorithms in forestry between 2016 and 2020. The 440 papers reviewed allowed to identify the most used algorithms in this period. The most widely used algorithms for soil filtering previously identified in the review were compared under different forest structural complexity (NEON sites). Three-point densities (±20, 8 and 1 p/m2) were considered to check which algorithm was more suitable for each NEON site and their overall performance. Out of all algorithms tested, Cloth Simulation Filter (CSF; MMCE = 0.967) and Progressive Triangulated Irregular Network (PTIN; MMCE = 0.951), outperformed the rest at all study sites.

We demonstrate a method, to detect shadows and to remove them to improve the quality in order to obtain better results in land cover classification. For shadow mask detection, specific color channels of L*a*b color space and H-S-I color space are selected and are processed by two shadow detection methods: Particle Swarm Optimization and the spectral H-L ratio. Shadow removal is performed in the Y-cb-cr color space adjusting the color values in shadowy regions. This adjusted Y-cb-cr image is transformed into the RGB image which is used for land cover classification and compared to classification results achieved with shadowed RGB images.

Accurate mapping of maize crops is crucial for sustainable agriculture and food security in developing countries. This study explores the application of machine learning algorithms, specifically Random Forest (RF) and k-Nearest Neighbors (kNN), to map maize crops in South Africa's Free State province. Using optical Sentinel-2 and Sentinel-1 radar images, the RF algorithm achieves an impressive 90% accuracy, while the kNN algorithm demonstrates a notable 86% precision. These findings provide valuable information for decision-makers and agronomists in selecting optimal mapping approaches. Future studies can build on these insights to refine strategies to address food insecurity and advance sustainable agriculture.

This work summarizes the progress and results of one of the strategic initiatives of the Italian Space Agency, the "Innovation for Downstream Preparation program" (I4DP), which aims to promote the use of remote sensing technologies to increase the resilience of urban environments and communities. Projects will be presented for the development of services and applications based on satellite data that support the processes of territorial management and protection of its resources, such as infrastructure monitoring, sustainable urban planning, and the fight against the effects of climate change.

The TUS:CAN project, in agreement between the Italian Space Agency (ASI), Tuscany Region, and the University of Siena, aims to define a step-by-step approach for carrying out land cover mapping of the artificial materials of man-made features in urban areas. Sentinel-2 data have been processed for natural/artificial mapping by Google Earth Engine (GEE) using the Random Forest (RF) classifier. The binary classification output serves as the mask layer to identify man-made surfaces that have been successively categorized in terms of artificial materials using PRISMA hyperspectral data. Spectral signatures have been collected as ground truths using an ASD FieldSpec 3 portable spectroradiometer for various typical artificial materials in urban areas. The PRISMA data cube has been used to assess the percentages of different types of artificial material covering roofs and pavements. Accuracy assessment of the classification has been carried out using aerial hyperspectral data cubes from Itres CASI-1500 and SASI-600 sensors.

MIDS (Monitoring of Water Infrastructures with Satellite Data) aims to enhance skills in satellite image processing, consolidating solutions for Monitoring of Irrigation and Drainage Infrastructures managed by consortia and basin authorities. It focuses on applying satellite data to control and monitor the efficiency and stability of irrigation networks, enhancing land value, production competitiveness, agricultural income, and employment. MIDS fills gaps in current monitoring systems using Multispectral, Hyperspectral, and SAR data. It develops three innovative products for Reclamation Consortia: Leaks monitoring in irrigation networks, Infrastructural monitoring of dams, and Pollutant detection in discharge channels. The test and validation area is the Reclamation Consortium of the Lower Volturno Basin (Italy), covering over 186,967 Ha

Planetek Italia, with the support of GAP, is developing an operational service build upon the existing Rheticus® Safeway, developed by Planetek within the Horizon 2020 Safeway project concluded in February 2022. The initiative, supported by the Italian Space Agency (ASI) under the I4DP Market project, targets providing a comprehensive solution aligned with the guidelines for risk classification, safety assessment, and monitoring of existing bridges, as outlined by the Italian Ministry of Infrastructure and Transport. The operational infrastructure monitoring service supports the predictive maintenance of roads, railways, and bridges.

GARMOSAT project aims to integrate EO satellites data of the ESA Copernicus Constellation and ASI PRISMA satellites, with ground measurements of GHG (focus on methane) in Municipal Solid Waste landfills of Latium Region without capping. Other ancillary results will be comparison analyses between Sentinel 1A SAR data and UAS data to create and validate a digital model of the landfill. Results of these analyses will be useful to build a verified and affordable “model” integrated in an operational SW visual tool to support decision makers of civil protection and authorities. In this project, we use and compare several detection methods, in particular field analysis and specific sensors as UAS payloads for ground truth, satellite data for GHG leak detection over large landfill sites and meteorological data to build an affordable operational model. The quantification of the GHG content will happen through chemical composition analysis of biogas produced by landfills (LFG) from punctual samples acquisition. We plan to employ a parallel approach using satellite data from Sentinel-2, Sentinel-3, Sentinel-5P and ASI PRISMA data, to see if alterations are correlated with ground truth data.

Health monitoring of transport infrastructure is crucial for efficient management and safety, particularly in the wake of recent global bridge collapses. This is the case of italy, where aligning with European guidelines, specific protocols for bridge assessment and monitoring were issued. While traditional visual inspections are accurate, they suffer from low repeatability and high costs. Recent advancements in Non-Destructive Testing (NDT), especially LiDAR technologies, offer promising solutions. This study focuses on utilizing Terrestrial Laser Scanner (TLS) point clouds for automated defect identification, emphasizing signal amplitude for enhanced accuracy. By clustering point clouds and integrating signal amplitude, structural anomalies like cracks and corrosion can be precisely identified. Experimental results demonstrate the effectiveness of this approach in reducing inspection time and streamlining the workflow, paving the way for more efficient monitoring of transport infrastructure.

Masonry bridges are heritage assets that have been historically under long service time. Preserving users’ safety and ensuring continuous service require the implementation of robust structural health monitoring (SHM) programs. Remote sensing, specifically the ground-based interferometric radar (GBIR), has proven effective in monitoring bridge structures in static and dynamic modes. However, to reach proper accuracy in target location, system control for the radar beam of radiation is crucial. As the piers are the bridge supporting systems and set the boundary conditions for these structures, they directly affect the structure’s vibration and natural frequency. In this paper, the dynamic behaviour of a masonry bridge pier was investigated by integration of GBIR and Augmented Reality (AR). AR allowed real-time system control by projecting the beam of radiation directly on to the structure. Data were investigated using multi-dimensional signal processing techniques for feature extraction. Results show that AR is key for integration with GBIR in real-time monitoring of bridges.

Urban Green infrastructure is an essential part of the urban ecosystem and is a sink for extreme heat and various emissions. Urbanization has been developing alarmingly due to a tall and densely built environment. The effect of green infrastructure in combating the new development should be quantified to balance the phenomena. Trees cover most green urban infrastructures, such as parks, woodlands, and streets. Hence, the essential characteristics of these trees, including tree height, canopy area, and tree health, should be monitored to understand the density and distribution of overall urban green cover. This monitoring has two benefits: primarily, it provides an understanding of the existing situation of the trees, and secondly, it helps in the assessment their impact during extreme weather conditions. Existing urban tree inventories and monitoring schemes are based on spatial sampling assessment techniques and visual inspections, which are limited in space and time. Remote sensing data provides prominent tools for quantifying green infrastructure and identifying changes over time. In this study, multi-source remote sensing datasets such as LiDAR and satellites are used to genera

Two different methodological approaches for the investigation of the Surface Urban Heat Island (SUHI) phenomenon in Italian regional capitals, based on the characterization of the municipal territories based on the Real Estate Market Observatory of the National Revenue Agency of Italy, and on the urbanization level of the Institute for Environmental Protection and Research of Italy . The obtained results and the WebGIS tool developed in the MIRIFICUS project provide useful information to Public Administrations for planning SUHI mitigation interventions.

In recent years, considerable attention has been directed towards the management of transportation infrastructure by private stakeholders and public institutions, aimed at adopting more effective and efficient technologies for monitoring and maintenance purposes. This research explores the integration of information derived from satellite data, proposing a novel methodology for managing multi-sensor survey data by integrating satellite remote sensing into a dynamic Digital Twin of an infrastructure. To achieve this, Synthetic Aperture Radar (SAR) data were processed using Multi-Temporal SAR Interferometry (MT-InSAR) to detect potential damages and changes in transport infrastructure and their surrounding environment.

UNESCO Global Geoparks are considered as territories with natural cultural heritage sites of exceptional geological importance. Nowadays, remote sensing has proven to be an effective, cost-efficient and non-invasive tool for the monitoring, and thus protection and management of such areas, contributing to their sustainable development. In this framework, Unmanned Aerial Vehicle (UAV) and Terrestrial Laser Scanning (TLS) data were utilized to map the geosite of Chelmos – Vouraikos UNESCO Global Geopark “Styx waters”. The main objective of the research was the digitalization of the geosite in order to create a digital counterpart, easily accessible to people who do not have the opportunity to visit it physically. 3D UAV and TLS representations of the geosite were generated.

Neural Radiance Field (NeRF) and photogrammetry are two methods that are used to replicate real-life objects and environments to produce high-quality models. This paper examines the use of NeRF and Photogrammetry techniques for documenting cultural heritage. To compare and examine the capabilities of both techniques, the documentation of the 14th-century church of Panagia Karmiotissa located in the Limassol region in Polemidia was utilized. The church was originally a Carmelite monastery from the 14th century A.D. Today, only the church exists, which was restored by the Department of Antiquities in 2001, with the name of Panagia Karmiotissa. On the north side of the church are the ruins of the monastery complex that are reduced to heaps of stones. The church of Panagia Karmiotissa is the only preserved Gothic church in the Limassol area during the time of the Frankish House of Lusignan in Cyprus. The church and Monastery area have been declared as an ancient monument by the Antiquities Department.

Diverse archaeological activities such as archaeological site documentation, excavation planning or cultural heritage management can benefit from 3D metric data produced by Unmanned Aerial Vehicles or Terrestrial Laser scanners. In the current study, 3D point clouds from three different commercial UAVs and one TLS are compared. Specifically, two tetracopters and an hexacopter were used to map the Ancient Helike site. At the same time, a Terrestrial Laser Scanner (BLK 360) was also used to collect 3D point clouds on the ground level. Hundreds of ground control points collected with RTK GNSS sensor were used to validate the accuracy of the derived point clouds. The validation procedure was performed in ArcMap through the comparison of 3D representations with the collected GNSS measurements.

The monitoring of cultural heritage sites using digital twins provides a dynamic visualization of the sites and monitors changes resulting from natural hazards and climate change. Digital twins can also be used as a tool for documentation, monitoring and management of cultural heritage sites by providing information about the current status and condition of the site.Various case studies were examined within this paper to examine how digital twins can be applied for monitoring and documentation for cultural heritage. One of the case studies in this paper is within the TRIQUETRA project, that studies the effects of climate change and natural hazards on cultural heritage and remediation using state-of-the-art techniques. Choirokoitia is a UNESCO World Heritage Site and is one of the best-preserved Neolithic sites in the Mediterranean. Through the TRIQUETRA project, Choirokoitia, Cyprus is used as one of the pilot studies. Another case study is the Agia Karmiotissa church, which is funded under the EXCELSIOR project where digital twins were used on a 14th-century church to document and monitor the site during the archaeological excavations.

Archaeological sites that are submerged in shallow waters are subject to various environmental threats, including anthropogenic factors, climate change and environmental conditions. Due to their archaeological significance, they are vulnerable to extreme risks from deterioration due to land deformation, flooding, acid rain, erosion, and man-made hazards like illegal excavations and tourist activities. Such threats not only endanger the structural integrity of these monuments, sometimes resulting to total destruction, and loss of cultural heritage and history. By integrating various Earth observation satellite images with and aerial imagery, the study aims to examine a methodology for the monitoring of underwater cultural heritage sites. This approach provides an understanding of the impacts of climate change as well as the human impact of various activities that affect the coastlines of cultural heritage sites and also provides a tool for developing proactive measures to safeguard heritage assets.