Статті (ММАД)
Постійне посилання зібрання
У зібранні розміщено матеріали, що опубліковані або готуються до публікації в наукових журналах та збірниках.
Переглянути
Перегляд Статті (ММАД) за Автор "Kussul, Nataliia"
Зараз показуємо 1 - 12 з 12
Результатів на сторінці
Налаштування сортування
Документ Відкритий доступ A generalized model for mapping sunflower areas using Sentinel-1 SAR data(Elsevier, 2024) Qadir, Abdul; Skakun, Sergii; Kussul, Nataliia; Shelestov, Andrii; Becker-Reshef, InbalExisting crop mapping models, rely heavily on reference (calibration) data obtained from remote sensing observations. However, the transferability of such models in space and time, without the need for additional extensive datasets remains a significant challenge. There is still a large gap in developing generalized classification models capable of mapping specific or multiple crops with minimal calibration data. In this study, we present a generalized automatic approach for sunflower mapping at 20-m spatial resolution, using the C-band Sentinel-1 (S1) synthetic aperture radar (SAR) data driven by previously developed phenological metrics. These metrics characterize the directional behavior of the sunflower head, capturing distinct backscattering responses in SAR data acquired from ascending and descending orbits. Specifically, we utilize SAR-derived backscatter values in VH and VV polarization, as well as their ratio VH/VV, as input features to a random forest classifier that was calibrated for the year 2022 in Ukraine. This model is further directly applied to selected sites for multiple years in Ukraine (generalization in time) and other major sunflower producing countries (generalization in space): Ukraine for 2018–2020, and Hungary, France, Russia and USA for 2018. Our results reveal that the model based on features acquired from descending orbits outperforms its ascending orbit counterpart because of the directional behavior of sunflower: user's accuracy (UA) of 96%, producer's accuracy (PA) of 97% and F-score of 97% (descending) compared to UA of 90%, PA of 89% and F-score of 90% (ascending). When generalized to other years and countries, our model achieves an F-score exceeding 77% for all cases, with the highest F-scores (>91%) observed in Ukraine and Russia sites and the lowest (77%) for the US site. We further utilize the produced maps (pixel-based) for the selected regions and years to estimate sunflower planted areas using a statistical sampling-based approach. Our estimates yield the relative root mean square error (RMSE) as 19.7% of the mean area, when compared to the reference data from official statistics and reference maps. These findings demonstrate the robustness of our proposed approach across space and time in generating accurate sunflower maps, its ability to mitigate cloud cover issues through spaceborne SAR data acquisitions, and its potential for obtaining estimates of sunflower planted areas. This research emphasizes the importance of developing interpretable and domain-specific machine learning models that can be readily extended to multiple geographical regions with little to no labelled datasets.Документ Відкритий доступ Air Quality Estimation in Ukraine Using SDG 11.6.2 Indicator Assessment(Remote Sensing, 2021-11) Shelestov, Andrii; Yailymova, Hanna; Yailymov, Bohdan; Kussul, NataliiaДокумент Відкритий доступ Assessing damage to agricultural fields from military actions in Ukraine: An integrated approach using statistical indicators and machine learning(2023) Kussul, Nataliia; Drozd, Sofiia; Yailymova, Hanna; Shelestov, Andrii; Lemoine, Guido; Deininger, KlausThe ongoing full-scale Russian invasion of Ukraine has led to widespread damage of agricultural lands, jeopardizing global food security. Timely detection of impacted fields enables quantification of production losses, guiding recovery policies and monitoring military actions. This study presents a robust methodology to automatically identify agricultural areas damaged by wartime ground activities using free Sentinel-2 satellite data. The 10 m resolution spectral bands and vegetation indices are leveraged, alongside their statistical metrics over time, as inputs to a Random Forest classifier. The algorithm efficiently pinpoints damaged fields, with accuracy metrics around 0.85. Subsequent anomaly detection delineates damages within the fields by combining spectral bands and indices. Applying the methodology over 22 biweekly periods in 2022, approximately 500 thousand ha of cropland across 10 regions of Ukraine were classified as damaged, with the most significant impacts occurring from March to September. The algorithm provides updated damage information despite cloud cover and vegetation shifts. The approach demonstrates the efficacy of automated satellite monitoring to assess agricultural impacts of military actions, supporting recovery analysis and documentation of war crimes.Документ Відкритий доступ Biophysical Impact of Sunflower Crop Rotation on Agricultural Fields(Sustainability, 2022) Kussul, Nataliia; Deininger, Klaus; Shumilo, Leonid; Lavreniuk, Mykola; Ayalew Ali, Daniel; Nivievskyi, OlegДокумент Відкритий доступ Discrete Atomic Transform-Based Lossy Compression of Three-Channel Remote Sensing Images with Quality Control(Remote Sensing, 2021) Makarichev, Victor; Vasilyeva, Irina; Lukin, Vladimir; Vozel, Benoit; Shelestov, Andrii; Kussul, NataliiaДокумент Відкритий доступ Fire Danger Assessment Using Moderate-Spatial Resolution Satellite Data(MDPI, 2023) Kussul, Nataliia; Fedorov, Oleh; Yailymov, Bohdan; Pidgorodetska, Liudmyla; Kolos, Liudmyla; Yailymova, Hanna; Shelestov, AndriiFire is one of the most common disturbances in natural ecosystems. The analysis of various sources of information (official and unofficial) about the fires in Ukraine (2019–2020) showed a lack of timely and reliable information. Satellite observation is of crucial importance to provide accurate, reliable, and timely information. This paper aims to modify the index of fire danger of a forest’s FWI by increasing its precision, based on the use of higher spatial resolution satellite data. A modification of the FWI method involves the utilization of the soil moisture deficit, in addition to the six subindices of the FWI system. In order to calculate the subindices values, weather data from the Copernicus Atmosphere Monitoring Service were used. Soil moisture deficit is calculated using Sentinel-1 radar satellite data on the water saturation degree of the soil surface layer and geospatial parameters from the 3D Soil Hydraulic Database of Europe. The application of the proposed methodology using the specified satellite, weather, and geospatial data makes it possible to assess fire danger on a continental scale with a spatial resolution of 250 m, 1 km, and a daily temporal resolution. Validation of the proposed method for modifying the FWI system demonstrates an improvement in the precision and relevance of fire danger prediction.Документ Відкритий доступ Generative adversarial network augmentation for solving the training data imbalance problem in crop classification(2023) Shumilo, Leonid; Okhrimenko, Anton; Kussul, Nataliia; Drozd, Sofiia; Shkalikov, OlehDeep learning models offer great potential for advancing land monitoring using satellite data. However, they face challenges due to imbalanced real-world data distributions of land cover and crop types, hindering scalability and transferability. This letter presents a novel data augmentation method employing Generative Adversarial Neural Networks (GANs) with pixel-to-pixel transformation (pix2pix). This approach generates realistic synthetic satellite images with artificial ground truth masks, even for rare crop class distributions. It enables the creation of additional minority class samples, enhancing control over training data balance and outperforming traditional augmentation methods. Implementing this method improved the overall map accuracy (OA) and intersection over union (IoU) by 1.5% and 2.1%, while average crop type classes’ user accuracy (UA) and producer accuracies (PA), as well as IoU, were improved by 11.2%, 6.4% and 10.2%.Документ Відкритий доступ Improving spatial resolution of chlorophyll-a in the Mediterranean sea based on machine learning(ХАІ, 2024) Yailymov, Bohdan; Kussul, Nataliia; Henitsoi, Pavlo; Shelestov, AndriiThe objective of this study is to increase the spatial resolution of data on the level of chlorophyll-a in the Mediterranean Sea using satellite images and ground measurements. The goal of this study is to develop an information technology based on machine learning to create chlorophyll-a concentration maps with high spatial resolution for the pilot areas of the Mediterranean Sea. Traditional ground-based methods for measuring chlorophyll-a are time-consuming, expensive, and have limited spatial and temporal coverage. Therefore, satellite observations have become an effective tool for monitoring chlorophyll-a over large areas. Low spatial resolution satellite data such as GCOM-C/SGLI and Sentinel-3 OLCI allow measurements of chlorophyll-a concentration at the sea surface. However, these data have limited accuracy and spatial resolution, which creates challenges for monitoring local changes in coastal zones and small water areas. Tasks: to analyze available satellite data and ground-based measurements of chlorophyll-a for the Mediterranean Sea; to investigate the correlation between satellite data of different spatial resolutions and ground measurements; to select informative features from satellite data for building machine learning models; and to develop models for increasing the spatial resolution of chlorophyll-a based on regression and machine learning algorithms. Obtained results: information technology combining satellite data with ground measurements in the Google Earth Engine cloud platform is proposed; correlations between satellite measurements of chlorophyll-a and ground data are investigated; models based on Random Forest and Multilayer Perceptron with coefficients of determination up to 0.36 and correlation of 0.6 with test data are built; chlorophyll-a maps with a spatial resolution of 10 m are created for the pilot area near Cyprus. Conclusions. The developed information technology allows the effective combination of satellite data of different spatial resolutions and ground measurements to increase the accuracy and detail of chlorophyll-a maps in the Mediterranean Sea. Further research involves improving the preprocessing of satellite data, using more features, involving data from other regions, and applying more sophisticated machine learning models.Документ Відкритий доступ Is Soil Bonitet an Adequate Indicator for Agricultural Land Appraisal in Ukraine?(Sustainability, 2021) Shumilo, Leonid; Lavreniuk, Mykola; Skakun, Sergii; Kussul, NataliiaДокумент Відкритий доступ Quantifying War-Induced Crop Losses in Ukraine in Near Real Time to Strengthen Local and Global Food Security(Elsevier Ltd, 2023) Deininger, Klaus; Ali, Daniel Ayalew; Kussul, Nataliia; Shelestov, Andrii; Lemoine, GuidoWe use a 4-year panel (2019–2022) of 10,125 village councils in Ukraine to estimate effects of the war started by Russia on area and expected yield of winter crops aggregated up from the field level. Satellite imagery is used to provide information on direct damage to agricultural fields; classify crop cover using machine learning; and compute the Normalized Difference Vegetation Index (NDVI) for winter cereal fields as a proxy for yield. Without conflict, winter crop area would have been 9.35 rather than 8.38 million ha, a 0.97 million ha reduction, only 14% of which can be attributed to direct conflict effects. The estimated drop associated with the conflict in NDVI for winter wheat, which is particularly pronounced for small farms, translates into an additional reduction of output by about 1.9 million tons for a total of 4.84 million tons. Taking area and yield reduction together suggests a war-induced loss of winter wheat output of up to 17% assuming the 2022 winter wheat crop was fully harvested.Документ Відкритий доступ Solar energy potential mapping in Ukraine through integration of GIS, remote sensing, and fuzzy logic(Associazione Italiana di Telerilevamento (AIT), 2024) Drozd, Sofiia; Kussul, NataliiaThe Green Deal strategic plan for the development of renewable energy until 2030 is ofparticular importance in the context of the restoration of Ukraine’s post-war energy infrastruc-ture. One of the key topics is the analysis of the possibilities of installing large solar powerplants in Ukraine. In this article, a multi-criteria analysis of the suitability of the territory ofUkraine is carried out on the basis of climatic, topographic and land use criteria. To assess landsuitability, criteria standardized using fuzzy logic with weights determined by experts throughthe method of pairwise comparisons were combined using a weighted sum model. Uponcompleting the study, a suitability map was generated, depicting zones with varying levels ofsuitability (ranging from 0 to 1) for solar power plant placement. It was found that more than35.68% of the country has average values of the suitability index (0.65–0.7), and approximately18.82% show high indicators (<0.7). Conditions are especially favorable in the south of Ukraine.Документ Відкритий доступ Transfer learning and single-polarized SAR image preprocessing for oil spill detection(Elsevier, 2025) Kussul, Nataliia; Salii, Yevhenii; Kuzin, Volodymyr; Yailymov, Bohdan; Shelestov, AndriiThis study addresses the challenge of oil spill detection using Synthetic Aperture Radar (SAR) satellite imagery, employing deep learning techniques to improve accuracy and efficiency. We investigated the effectiveness of various neural network architectures and encoders for this task, focusing on scenarios with limited training data. The research problem centered on enhancing feature extraction from single-channel SAR data to improve oil spill detection performance. Our methodology involved developing a novel preprocessing pipeline that converts single-channel SAR data into a three-channel RGB representation. The preprocessing technique normalizes SAR intensity values and encodes extracted features into RGB channels. Through an experiment, we have shown that a combination of the LinkNet with an EfficientNet-B4 is superior to pairs of other well-known architectures and encoders. Quantitative evaluation revealed a significant improvement in F1-score of 0.064 compared to traditional dB-scale preprocessing methods. Qualitative assessment on independent SAR scenes from the Mediterranean Sea demonstrated better detection capabilities, albeit with increased sensitivity to look-alike. We conclude that our proposed preprocessing technique shows promise for enhancing automatic oil spill segmentation from SAR imagery. The study contributes to advancing oil spill detection methods, with potential implications for environmental monitoring and marine ecosystem protection.