Sentinel-2 Imagery Research Articles

A Broadscale Assessment of Sentinel-2 Imagery and the Google Earth Engine for the Nationwide Mapping of Chlorophyll a

A Broadscale Assessment of Sentinel-2 Imagery and the Google Earth Engine for the Nationwide Mapping of Chlorophyll a

Extracting shallow water depth from the fusion of multi-temporal ICESat-2 data and multi-spectral imageries

With the development of satellite remote sensing and laser altimetry data, the fusion of laser altimetry data and multi-spectral images for shallow water depth inversion has become an economically convenient way. However, there are few methods that take into account the temporal dimension of data to integrate the results of water depth inversion in multiple temporal phases. In response to this issue, this article proposed a shallow water depth inversion method that integrates multi-source and multi-temporal remote sensing data. This method utilized the random forest (RF) algorithm to estimate the water depth values at different time, taking into account the contextual information of Sentinel-2 imagery and using the overall least squares as the theoretical model to fuse the multi-temporal water depth inversion results. This article took the Yongle Islands in the South China Sea as the research area and conducted the shallow water inversion experiments using ICESat-2 (Ice, Cloud and Land Elevation Satellite 2) data from 5 temporal phases and one Sentinel-2 imagery. The results show that the mean value of RMSE (root mean square error) and R2 (determination coefficient) of the proposed method using single temporal imagery was 1.53m and 0.7610, which outperformed the inversion accuracy of traditional methods that ignore image context information. The RMSE and R2 of the multi-temporal fusion model was 1.15m and 0.8622, which was 0.08m and 0.0199m higher than existing median filtering fusion method.

Extraction of Terraces in Hilly Areas from Remote Sensing Images Using DEM and Improved U-Net

To extract terraced fields in hilly areas on a large scale in an automated and high-precision manner, this paper proposes a terrace extraction method that combines the Digital Elevation Model (DEM), Sentinel-2 imagery, and the improved U-Net semantic segmentation model. The U-Net model is modified by introducing Attention Gate modules into its decoding modules to suppress the interference of redundant features and adding Dropout and Batch Normalization layers to improve training speed, robustness, and fitting ability. In addition, the DEM band is combined with the red, green, and blue bands of the remote sensing images to make full use of terrain information. The experimental results show that the Precision, Recall, F1 score, and Mean Intersection over Union of the proposed method for terrace extraction are improved to other mainstream advanced methods, and the internal information of the terraces extracted is more complete, with fewer false positive and false negative results.

Assessment of the impact of utility-scale photovoltaics on the surrounding environment in the Iberian Peninsula. Alternatives for the coexistence with agriculture

The rapid growth of photovoltaic solar energy, to achieve decarbonization, has been accompanied by increasing land occupation and the subsequent concern in the agroforestry sector. The increase in land area occupied has been of 20% in recent years, boosting solar electricity production to 5.9% of the total in Europe. This fact raises the question of the impact on vegetation greenness and moisture in the rural environment, something that has not always been considered.Image analysis is presented as one of the most effective tools to estimate the variation of vegetation greenness and moisture. For this, terrestrial images, like those from unmanned aerial vehicles, can be used; however, this limits the amount of information available and/or increases the cost. The use of satellite images in different bands is a relatively new tool that can be exploited for the analysis of solar plants impact. This work presents a new way to use Sentinel imagery to analyse the impact of utility-scale solar plants on vegetation and moisture of the surrounding areas. According to our results, a moderate decrease in weighted index for both moisture (5%) and vegetation (3%) occurred after solar plant installation. It is expected that these results can be of help for the design of new PV and agrivoltaic plants, originating the Ground-Integrated Photovoltaics (GIPV).

Landslide extraction using a novel empirical method and binary semantic segmentation U-NET framework using sentinel-2 imagery

ABSTRACT Artificial intelligence (AI) has achieved a remarkable place in solving complex problems in almost all disciplines. Based on the recent notable performances of machine learning and deep learning techniques for rapid and automatic landslide identifcation, it is observed that availability of quality training data, proper model training and associated cost are crucial for developing such frameworks. Therefore, the primary objective of the study is to propose a novel empirical algorithm, DvD, for rapid landslide identification using Sentinel-2 imagery and comparatively evaluate its performance to a deep learning architecture popularly used in feature extraction problems, binary semantic segmentation U-NET (BSS-UNET) framework. The empirical method has been investigated over a dataset diverse in topography and land cover to evaluate its efficacy. The proposed BSS-UNET framework is trained on the landslide database provided by the Institute of Advance Research in Artificial Intelligence (IARAI) in Landslide4Sense 2022 challenge which achieved a high mIoU value of 0.78 with 84.23% precision, 65% recall and 73.32 F1-score. The DvD algorithm outperformed the BSS-UNET framework and achieved 0.80 mIoU when applied to the IARAI dataset. The proposed empirical method has the potential to serve as large-scale rapid landslide inventory preparation subject to the availability of cloud-free satellite imagery.

Automated Built-Up Infrastructure Land Cover Extraction Using Index Ensembles with Machine Learning, Automated Training Data, and Red Band Texture Layers

Automated built-up infrastructure classification is a global need for planning. However, individual indices have weaknesses, including spectral confusion with bare ground, and computational requirements for deep learning are intensive. We present a computationally lightweight method to classify built-up infrastructure. We use an ensemble of spectral indices and a novel red-band texture layer with global thresholds determined from 12 diverse sites (two seasonally varied images per site). Multiple spectral indexes were evaluated using Sentinel-2 imagery. Our texture metric uses the red band to separate built-up infrastructure from spectrally similar bare ground. Our evaluation produced global thresholds by evaluating ground truth points against a range of site-specific optimal index thresholds across the 24 images. These were used to classify an ensemble, and then spectral indexes, texture, and stratified random sampling guided training data selection. The training data fit a random forest classifier to create final binary maps. Validation found an average overall accuracy of 79.95% (±4%) and an F1 score of 0.5304 (±0.07). The inclusion of the texture metric improved overall accuracy by 14–21%. A comparison to site-specific thresholds and a deep learning-derived layer is provided. This automated built-up infrastructure mapping framework requires only public imagery to support time-sensitive land management workflows.

Automated Mapping of Land Cover Type within International Heterogenous Landscapes Using Sentinel-2 Imagery with Ancillary Geospatial Data.

A near-global framework for automated training data generation and land cover classification using shallow machine learning with low-density time series imagery does not exist. This study presents a methodology to map nine-class, six-class, and five-class land cover using two dates (winter and non-winter) of a Sentinel-2 granule across seven international sites. The approach uses a series of spectral, textural, and distance decision functions combined with modified ancillary layers (such as global impervious surface and global tree cover) to create binary masks from which to generate a balanced set of training data applied to a random forest classifier. For the land cover masks, stepwise threshold adjustments were applied to reflectance, spectral index values, and Euclidean distance layers, with 62 combinations evaluated. Global (all seven scenes) and regional (arid, tropics, and temperate) adaptive thresholds were computed. An annual 95th and 5th percentile NDVI composite was used to provide temporal corrections to the decision functions, and these corrections were compared against the original model. The accuracy assessment found that the regional adaptive thresholds for both the two-date land cover and the temporally corrected land cover could accurately map land cover type within nine-class (68.4% vs. 73.1%), six-class (79.8% vs. 82.8%), and five-class (80.1% vs. 85.1%) schemes. Lastly, the five-class and six-class models were compared with a manually labeled deep learning model (Esri), where they performed with similar accuracies (five classes: Esri 80.0 ± 3.4%, region corrected 85.1 ± 2.9%). The results highlight not only performance in line with an intensive deep learning approach, but also that reasonably accurate models can be created without a full annual time series of imagery.

Spatial Insights into Drought Severity: Multi-Index Assessment in Małopolska, Poland, via Satellite Observations

This study focuses on the assessment of drought severity, employing a comparative analysis between the normalized multi-band drought index (NMDI; calculated using Sentinel-2 imagery) and the combined drought indicator (CDI). The research aims to pinpoint the most accurate and reliable remote sensing techniques, which reflect ground-based measurements, thereby enhancing the precision of drought monitoring systems. By investigating the specific area of Małopolska, the study not only contributes to the global knowledge base on drought assessment methodologies but also addresses the regional needs for improved drought management practices. Through an exhaustive analysis correlating satellite-derived indices with in situ meteorological data, this research elucidates the potential of integrating NMDI and CDI for a comprehensive understanding of drought dynamics in Małopolska. In particular, the indices depict different dynamics of drought levels, as well as the location of regions more prone to its occurrence. The findings are poised to advance drought monitoring capabilities, offering significant insights for agricultural sustainability and water resource management in the region.

Monitoring Spatial–Temporal Variations in River Width in the Aral Sea Basin with Sentinel-2 Imagery

Rivers in arid regions serve as crucial freshwater resources for local communities and play an essential role in global hydrological and biogeochemical cycles. The Aral Sea Basin (ASB) in Central Asia is characterized by an arid climate and river dynamics that are sensitive to climate change and human activities. Monitoring the spatiotemporal variations in river water extent in the ASB is essential to maintain an ecological balance and ensure water security. In this study, we extracted data regarding monthly river water bodies in the ASB from 2017 to 2022 by synthesizing monthly Sentinel-2 images. The water extents on the Sentinel images were automatically mapped using the Otsu method, and the river widths for all river channels were calculated using the RivWidth algorithm. We investigated the relationships between the river dynamics and the geomorphology, climatic change, human activities, and the annual and interannual variations in the river width in different reaches of the basin. The results show a seasonal variability in the river width, with most rivers reaching the largest width in the warm season and a few rivers in the middle and lower areas reaching the valley value in the warm season. Compared to their tributaries, the mainstem in the middle/lower regions showed less seasonal variability. According to interannual analysis, most of the rivers in the ASB significantly narrowed between 2017 and 2022, a phenomenon which is generally impacted by temperature and evapotranspiration variations. Comparisons show that our results provide improved information about the narrow river reaches and denser river networks compared to the previous global dataset, demonstrating the advantageous properties of high spatial resolution in Sentinel-2 imagery.

National tree species mapping using Sentinel-1/2 time series and German National Forest Inventory data

Spatially explicit and detailed information on tree species composition is critical for forest management, nature conservation and the assessment of forest ecosystem services. In many countries, forest attributes are monitored regularly through sample-based forest inventories. In combination with satellite imagery, data from such forest inventories have a great potential for developing large-area tree species maps. Here, the high temporal resolution of Sentinel-1 and Sentinel-2 has been useful for extracting vegetation phenology, information that may also be valuable for improving forest tree species mapping.The objective of this study was to map the main tree species in Germany using combined Sentinel-1 and Sentinel-2 time series, and to identify and address challenges related to the use of National Forest Inventory (NFI) data in remote sensing applications. We generated cloud free time series with 5-day intervals from Sentinel-2 imagery and combine those with monthly Sentinel-1 backscatter composites. Further, we incorporate information on topography, meteorology, and climate to account for environmental gradients. To use NFI data for training machine learning models, we address the following challenges: 1) link satellite pixels with variable radius NFI plots, for which the precise area is unknown, and 2) efficiently utilize mixed-species NFI plots for model training and validation. In the past, accuracies for pixel-level species maps were often estimated solely for homogeneous pure-species stands. In this study, we assess how well pixel-level maps generalize to mixed plot conditions.Our results show the potential of combined Sentinel-2 and Sentinel-1 time series with NFI data for tree species mapping in large, environmentally diverse landscapes. Classification accuracy in pure stands ranged between 72% and 97% (F1-score) for five dominant species, while mapping less frequent species remained challenging. When including mixed forest stands in the accuracy assessment, accuracy decreased by 4–14 percentage points for the most dominant species groups. Our study highlights the importance of including mixed-forest stands when training and validating tree species maps. Based on these results, we discuss potentials and remaining challenges for tree species mapping at the national level. Our findings allow to further improve national-level tree species mapping with medium to high resolution data and provide guidance for similar approaches in other countries where ground-based inventory data are available.

Sugarcane yield estimation in Thailand at multiple scales using the integration of UAV and Sentinel-2 imagery

Timely and accurate estimates of sugarcane yield provide valuable information for food management, bio-energy production, (inter)national trade, industry planning and government policy. Remote sensing and machine learning approaches can improve sugarcane yield estimation. Previous attempts have however often suffered from too few training samples due to the fact that field data collection is expensive and time-consuming. Our study demonstrates that unmanned aerial vehicle (UAV) data can be used to generate field-level yield data using only a limited number of field measurements. Plant height obtained from RGB UAV-images was used to train a model to derive intra-field yield maps based on 41 field sample plots spread over 20 sugarcane fields in the Udon Thani Province, Thailand. The yield maps were subsequently used as reference data to train another model to estimate yield from multi-spectral Sentinel-2 (S2) imagery. The integrated UAV yield and S2 data was found efficient with RMSE of 6.88 t/ha (per 10 m × 10 m pixel), for average yields of about 58 t/ha. The expansion of the sugarcane yield mapping across the entire region of 11,730 km2 was in line with the official statistical yield data and highlighted the high spatial variability of yields, both between and within fields. The presented method is a cost-effective and high-quality yield mapping approach which provides useful information for sustainable sugarcane yield management and decision-making.

Giant Aufeis in the Pangong Tso Basin: Inventory of a Neglected Cryospheric Component in Eastern Ladakh and Western Tibet

Cryosphere studies in High Mountain Asia (HMA) typically focus on glaciers, seasonal snow cover, and permafrost. As an additional and mostly overlooked cryosphere component, aufeis occurs frequently in cold-arid regions and covers extensive areas of the Trans-Himalaya and Tibetan Plateau. This largely neglected cryosphere component generally forms in winter from repeated freezing of seepage or overflow. In this article, the occurrence of aufeis fields in the endorheic Pangong Tso Basin (PTB), with a total area of 31,000 km2, is inventoried and examined. Based on a semi-automatic remote sensing approach using Sentinel-2 imagery, about 1000 aufeis fields were detected in the spring of 2019, covering a total area of approximately 86 km2 and with an average individual size of 0.08 km2, while the largest field covered an area of 14.8 km2. A striking contrast between the northern and southern portions of the PTB characterized the spatial distribution of large aufeis fields. All large (>0.5 km2) and 13 persisting aufeis fields were located along broad valleys in the northern portion. Furthermore, a multi-temporal comparison between 1994 and 2023 shows that the number of remaining aufeis fields in autumn varied between 8 and 29, with a maximum in 2019. Their total area ranged between about 0.3 km2 in 1994 and 2023 to about 1.2 km2 in 2015 and 2019. This study complements recent aufeis inventories from the Trans-Himalayan region of Ladakh and closes the gap to the Tibetan Plateau.

Comparative analysis of multi-source data for machine learning-based LAI estimation in Argania spinosa

In this study, we conducted a comprehensive assessment of Leaf Area Index (LAI) estimation using three distinct sources of satellite data: Sentinel-2 imagery, drone imagery (UAVs), and Mohammed VI satellite data. The main objective was to identify the most reliable and precise dataset for predicting LAI, with a focus on evaluating the performance of Random Forest models. For Sentinel-2 imagery, our Random Forest model achieves a robust R-squared (R2) value of 0.89, signifying a strong alignment between predicted and measured LAI values. The associated root-mean-square error (RMSE) is 0.4, indicating high predictive accuracy. In the context of UAVs, our Random Forest model excels, exhibiting an impressive R2 value of 0.93, highlighting a substantial correlation between predicted and measured LAI. The RMSE for drone imagery stands at 0.37, showcasing exceptional predictive accuracy. Finally, the Random Forest model trained on Mohammed VI satellite data yields an R2 value of 0.92, underlining its strong fit with measured LAI values. The RMSE for Mohammed VI imagery is 0.39, further underscoring the model's exceptional predictive accuracy. This comparative analysis underscores the importance of selecting the most suitable satellite data source for LAI estimation in Argania spinosa. UAV imagery emerges as the most accurate choice, closely followed by Mohammed VI Satellite and Sentinel-2 imagery. These findings offer valuable insights for effective monitoring of Argania spinosa and advancing sustainable land management practices in rural ecosystems.

Mapping Paddy Rice in Rice–Wetland Coexistence Zone by Integrating Sentinel-1 and Sentinel-2 Data

Accurate mapping of vegetation in the coexisting area of paddy fields and wetlands plays a key role in the sustainable development of agriculture and ecology, which is critical for national food security and ecosystem balance. The phenology-based rice mapping algorithm uses unique flooding stages of paddy rice, and it has been widely used for rice mapping. However, wetlands with similar flooding signatures make rice extraction in rice–wetland coexistence challenging. In this study, we analyzed phenology differences between rice and wetlands based on the Sentinel-1/2 data and used the random forest algorithm to map vegetation in the Poyang Lake Basin, which is a typical rice–wetland coexistence zone in the south of China. The rice maps were validated with reference data, and the highest overall accuracy and Kappa coefficient was 0.94 and 0.93, respectively. First, monthly median composited and J-M distance methods were used to analyze radar and spectral data in key phenological periods, and it was found that the combination of the two approaches can effectively improve the confused signal between paddy rice and wetlands. Second, the VV and VH polarization characteristics of Sentinel-1 data enable better identification of wetlands and rice. Third, from 2018 to 2022, paddy rice in the Poyang Lake Basin showed the characteristics of planting structure around the Poyang Lake and its tributaries. The mudflats were mostly found in the middle and northeast of Poyang Lake, and the wetland vegetation was found surrounding the mudflats, forming a nibbling shape from the lake’s periphery to its center. Our study demonstrates the potential of mapping paddy rice in the rice–wetland coexistence zone using the combination of Sentinel-1 and Sentinel-2 imagery, which would be beneficial for balancing the changes between paddy rice and wetlands and improving the vulnerability of the local ecological environment.

Temporal Analysis of Mangrove Forest Extent in Restoration Initiatives: A Remote Sensing Approach Using Sentinel-2 Imagery

The significance of mangroves and the associated risks they face have prompted government and the private sector to invest in projects aimed at conserving and restoring mangroves. Despite this interest, there is currently little information available on the effectiveness of these investments in mangrove conservation and restoration efforts. Therefore, this study aimed to use Sentinel-2 imagery with 10-m resolution through the Google Earth Engine to evaluate the effectiveness of these projects in mangrove areas in two regions: the Tahiry Honko project in Madagascar and the Abu Dhabi Blue Carbon Demonstration Project in the United Arab Emirates. The study compared the U-Net and SVM for mangrove classification. The U-Net model demonstrated superior performance, achieving an accuracy of 90%, with a Kappa coefficient value of 0.84. In contrast, the SVM had an overall accuracy of 86% and Kappa coefficient of 0.78. The analysis of changes in the mangrove area using U-Net model revealed a decline of 355 ha over four years in the Tahiry Honko project, while in the Abu Dhabi Project, the mangrove area increased by 5857 ha over 5 years. These findings can provide valuable information for policy-makers and management strategies.

A 10-m national-scale map of ground-mounted photovoltaic power stations in China of 2020

We provide a remote sensing derived dataset for large-scale ground-mounted photovoltaic (PV) power stations in China of 2020, which has high spatial resolution of 10 meters. The dataset is based on the Google Earth Engine (GEE) cloud computing platform via random forest classifier and active learning strategy. Specifically, ground samples are carefully collected across China via both field survey and visual interpretation. Afterwards, spectral and texture features are calculated from publicly available Sentinel-2 imagery. Meanwhile, topographic features consisting of slope and aspect that are sensitive to PV locations are also included, aiming to construct a multi-dimensional and discriminative feature space. Finally, the trained random forest model is adopted to predict PV power stations of China parallelly on GEE. Technical validation has been carefully performed across China which achieved a satisfactory accuracy over 89%. Above all, as the first publicly released 10-m national-scale distribution dataset of China’s ground-mounted PV power stations, it can provide data references for relevant researchers in fields such as energy, land, remote sensing and environmental sciences.

Index-based Approach in Relation to Built-up and LST Dynamics; A Study of Lahore, Pakistan

In 21st century, cities outpaced in size and also in density due to development of economic sector.Consequently, the wide spread expansion of urban areas is resulting in the loss of productive green cover and waterbodies. Therefore, realizing this alarming situation, the present study is aimed to investigate and evaluate the pattern ofurban expansion by considering two major land cover types (i) built-up area (ii) other classes (vegetation, waterbody,soil etc.) during the last six years (2015-2020). For this study Sentinel imagery was acquired from USGS EarthExplorer, while Modis Terra images were acquired from World View NASA. New built-up area index (NBUI),normalized difference vegetation index (NDVI), worldview water index (WV-WI) and land surface temperature (LST)were calculated in order to analyze variations in Lahore’s major land cover types and its varying temperature patterns.Spatial analysis presented the obvious impacts of land development on Lahore. NBUI indicated that the built-up areahas increased drastically from 34.0% in 2015 to 84.2% in 2020; NDVI analysis depicted a decline from 0.76% to0.73%, in the green spaces of Lahore during the study period; WV-WI portrayed inconsistent values of water bodies, agift of massive rise in the built-up area in Lahore. LST results presented that the temperature was 42.21°C in 2015,which simultaneously increased and recorded at 49.51°C in 2020. The increase in LST exhibited the alarming situationfor urban environment and can become threat to increase the air pollution level in Lahore. Therefore, this study willserve as a snapshot for policy makers to control the menace of unplanned urbanization by formulating stringent policiesto protect environment.

Plastic-covered greenhouses mapping in Morocco with Google Earth engine: Comparing Sentinel-2 and Landsat-8 data using pixel- and object-based methods

With the ever-increasing demand for food production, monitoring plastic-covered greenhouse (PCG) spatiotemporal distribution accurately holds significant importance for monitoring agricultural expansion and environmental impacts. For this purpose, in this paper, we used a random forest (RF) algorithm to produce multi-temporal PCGs maps in three different intensive agricultural Moroccan areas (Loukkos (LS), Gharb (GB), and Souss-Mass (SM)). Two classification approaches, object-based (OB) and pixel-based (PB), were applied to Landsat-8 (L8) and Sentinel-2 (S2) imagery via the Google Earth Engine (GEE) platform. Additionally, the RF performance is compared across five feature scenarios, combining spectral bands, spectral indices, and texture information. Further, we estimated the PCGs' spatiotemporal dynamics from 2018 to 2023. Findings indicated that the Blue and SWIR bands, Normalized Difference Tillage Index (NDTI), Plastic-Mulched Landcover Index (PMLI), variance, and sum average texture were the most significant variables that contributed to accuracy in PCGs classification. The 6-year average overall accuracy achieved >93.10% in mapping PCGs. Notably, the band combination “All” (Bands + Indices + Texture) produced high overall accuracy (>94.9%) and F1-score (>91.8%) compared to other feature combinations. Nevertheless, we found considerably improved classification accuracy by combining only spectral indices with texture (OA >94.6%). It was noticed that both OB and PB approaches reveal their similar performance in most cases, with slight differences. Besides, the PCGs estimated area analysis showed that SM area had the largest PCG coverage, approximately three times larger than the other areas, and it exhibited an overall growth of +8.45% between 2018 and 2023. Thus, this research demonstrates the RF and OB-PB's capability to accurately identify PCGs in GEE. In addition, these results can serve as valuable insights for policymakers and agricultural planners to efficiently monitor and manage PCGs expansion in intensive agricultural areas.

Automated spectral transfer learning strategy for semi-supervised regression on Chlorophyll-a retrievals with Sentinel-2 imagery

ABSTRACT Multispectral images make it possible to retrieve water quality parameters over a wide range and long time series by remote sensing. The robustness of traditional supervised machine learning models, which have been widely used, is affected in various regions, and these models are developed by in-situ measurements and limited satellite images. This study combined a spectral transfer learning strategy with a semi-supervised regression model to conduct data augmentation and developed a robust model for two inland lakes in Jilin Province, Northeast China, based on Sentinel-2 Multispectral Instruments images. Specifically, this study integrated an automated module for Balance Distribution Adaptation and a spectral features transfer method, and then applied it to Co-training Regressors for Chlorophyll-a retrievals to achieve the highest accuracy, which solved the selection of parameters and models. The Automated model was developed and achieved the best performance (R 2 is 0.91, the root mean square error is 2.62 μg L−1 and the mean absolute percentage error is 22.88%) compared with other models, and showed robustness in various lakes and months. Our results offer a reliable approach to provide an accurate Chlorophyll-a estimation of spatial–temporal variations.

Mangrove ecosystem species mapping from integrated Sentinel-2 imagery and field spectral data using random forest algorithm

Mangrove ecosystem species mapping from integrated Sentinel-2 imagery and field spectral data using random forest algorithm