A Burned Area Mapping Algorithm for Sentinel-2 Data Based on Approximate Reasoning and Region Growing

Sentinel-2 (S2) multi-spectral instrument (MSI) images are used in an automated approach built on fuzzy set theory and a region growing (RG) algorithm to identify areas affected by fires in Mediterranean regions. S2 spectral bands and their post- and pre-fire date (Dpost-pre) difference are interpreted as evidence of burn through soft constraints of membership functions defined from statistics of burned/unburned training regions; evidence of burn brought by the S2 spectral bands (partial evidence) is integrated using ordered weighted averaging (OWA) operators that provide synthetic score layers of likelihood of burn (global evidence of burn) that are combined in an RG algorithm. The algorithm is defined over a training site located in Italy, Vesuvius National Park, where membership functions are defined and OWA and RG algorithms are first tested. Over this site, validation is carried out by comparison with reference fire perimeters derived from supervised classification of very high-resolution (VHR) PlanetScope images leading to more than satisfactory results with Dice coefficient >0.84, commission error <0.22 and omission error <0.15. The algorithm is tested for exportability over five sites in Portugal (1), Spain (2) and Greece (2) to evaluate the performance by comparison with fire reference perimeters derived from the Copernicus Emergency Management Service (EMS) database. In these sites, we estimate commission error <0.15, omission error <0.1 and Dice coefficient >0.9 with accuracy in some cases greater than values obtained in the training site. Regression analysis confirmed the satisfactory accuracy levels achieved over all sites. The algorithm proposed offers the advantages of being least dependent on a priori/supervised selection for input bands (by building on the integration of redundant partial burn evidence) and for criteria/threshold to obtain segmentation into burned/unburned areas.

Download Full-text

A Fully Automatic, Interpretable and Adaptive Machine Learning Approach to Map Burned Area from Remote Sensing

ISPRS International Journal of Geo-Information ◽

10.3390/ijgi10080546 ◽

2021 ◽

Vol 10 (8) ◽

pp. 546

Author(s):

Daniela Stroppiana ◽

Gloria Bordogna ◽

Matteo Sali ◽

Mirco Boschetti ◽

Giovanna Sona ◽

...

Keyword(s):

Machine Learning ◽

Remote Sensing ◽

Region Growing ◽

Influential Factors ◽

Weighted Averaging ◽

Fire Season ◽

Burned Area ◽

Commission Error ◽

Burned Areas ◽

Fully Automatic

The paper proposes a fully automatic algorithm approach to map burned areas from remote sensing characterized by human interpretable mapping criteria and explainable results. This approach is partially knowledge-driven and partially data-driven. It exploits active fire points to train the fusion function of factors deemed influential in determining the evidence of burned conditions from reflectance values of multispectral Sentinel-2 (S2) data. The fusion function is used to compute a map of seeds (burned pixels) that are adaptively expanded by applying a Region Growing (RG) algorithm to generate the final burned area map. The fusion function is an Ordered Weighted Averaging (OWA) operator, learnt through the application of a machine learning (ML) algorithm from a set of highly reliable fire points. Its semantics are characterized by two measures, the degrees of pessimism/optimism and democracy/monarchy. The former allows the prediction of the results of the fusion as affected by more false positives (commission errors) than false negatives (omission errors) in the case of pessimism, or vice versa; the latter foresees if there are only a few highly influential factors or many low influential ones that determine the result. The prediction on the degree of pessimism/optimism allows the expansion of the seeds to be appropriately tuned by selecting the most suited growing layer for the RG algorithm thus adapting the algorithm to the context. The paper illustrates the application of the automatic method in four study areas in southern Europe to map burned areas for the 2017 fire season. Thematic accuracy at each site was assessed by comparison to reference perimeters to prove the adaptability of the approach to the context; estimated average accuracy metrics are omission error = 0.057, commission error = 0.068, Dice coefficient = 0.94 and relative bias = 0.0046.

Download Full-text

Hindcast and Near Real-Time Monitoring of Green Macroalgae Blooms in Shallow Coral Reef Lagoons Using Sentinel-2: A New-Caledonia Case Study

Remote Sensing ◽

10.3390/rs13020211 ◽

2021 ◽

Vol 13 (2) ◽

pp. 211

Author(s):

Maële Brisset ◽

Simon Van Wynsberge ◽

Serge Andréfouët ◽

Claude Payri ◽

Benoît Soulard ◽

...

Keyword(s):

Green Algae ◽

Algal Blooms ◽

New Caledonia ◽

Spectral Indices ◽

Green Macroalgae ◽

Spectral Bands ◽

Algae Blooms ◽

Trade Offs ◽

Sentinel 2

Despite the necessary trade-offs between spatial and temporal resolution, remote sensing is an effective approach to monitor macroalgae blooms, understand their origins and anticipate their developments. Monitoring of small tropical lagoons is challenging because they require high resolutions. Since 2017, the Sentinel-2 satellites has provided new perspectives, and the feasibility of monitoring green algae blooms was investigated in this study. In the Poé-Gouaro-Déva lagoon, New Caledonia, recent Ulva blooms are the cause of significant nuisances when beaching. Spectral indices using the blue and green spectral bands were confronted with field observations of algal abundances using images concurrent with fieldwork. Depending on seabed compositions and types of correction applied to reflectance data, the spectral indices explained between 1 and 64.9% of variance. The models providing the best statistical fit were used to revisit the algal dynamics using Sentinel-2 data from January 2017 to December 2019, through two image segmentation approaches: unsupervised and supervised. The latter accurately reproduced the two algal blooms that occurred in the area in 2018. This paper demonstrates that Sentinel-2 data can be an effective source to hindcast and monitor the dynamics of green algae in shallow lagoons.

Download Full-text

Accurate extraction of surface water in complex environment based on Google Earth Engine and Sentinel-2

PLoS ONE ◽

10.1371/journal.pone.0253209 ◽

2021 ◽

Vol 16 (6) ◽

pp. e0253209

Author(s):

Jianfeng Li ◽

Biao Peng ◽

Yulu Wei ◽

Huping Ye

Keyword(s):

Surface Water ◽

Google Earth ◽

Water Extraction ◽

Practical Significance ◽

Commission Error ◽

Complex Environment ◽

Visual Interpretation ◽

Water Index ◽

Google Earth Engine ◽

Sentinel 2

To realize the accurate extraction of surface water in complex environment, this study takes Sri Lanka as the study area owing to the complex geography and various types of water bodies. Based on Google Earth engine and Sentinel-2 images, an automatic water extraction model in complex environment(AWECE) was developed. The accuracy of water extraction by AWECE, NDWI, MNDWI and the revised version of multi-spectral water index (MuWI-R) models was evaluated from visual interpretation and quantitative analysis. The results show that the AWECE model could significantly improve the accuracy of water extraction in complex environment, with an overall accuracy of 97.16%, and an extremely low omission error (0.74%) and commission error (2.35%). The AEWCE model could effectively avoid the influence of cloud shadow, mountain shadow and paddy soil on water extraction accuracy. The model can be widely applied in cloudy, mountainous and other areas with complex environments, which has important practical significance for water resources investigation, monitoring and protection.

Download Full-text

A Deep Learning Approach for Burned Area Segmentation with Sentinel-2 Data

Remote Sensing ◽

10.3390/rs12152422 ◽

2020 ◽

Vol 12 (15) ◽

pp. 2422

Author(s):

Lisa Knopp ◽

Marc Wieland ◽

Michaela Rättich ◽

Sandro Martinis

Keyword(s):

Deep Learning ◽

Network Performance ◽

Near Infrared ◽

Remote Sensing Data ◽

Economic Consequences ◽

Validation Dataset ◽

Burned Area ◽

Burned Areas ◽

Spectral Bands ◽

Sentinel 2

Wildfires have major ecological, social and economic consequences. Information about the extent of burned areas is essential to assess these consequences and can be derived from remote sensing data. Over the last years, several methods have been developed to segment burned areas with satellite imagery. However, these methods mostly require extensive preprocessing, while deep learning techniques—which have successfully been applied to other segmentation tasks—have yet to be fully explored. In this work, we combine sensor-specific and methodological developments from the past few years and suggest an automatic processing chain, based on deep learning, for burned area segmentation using mono-temporal Sentinel-2 imagery. In particular, we created a new training and validation dataset, which is used to train a convolutional neural network based on a U-Net architecture. We performed several tests on the input data and reached optimal network performance using the spectral bands of the visual, near infrared and shortwave infrared domains. The final segmentation model achieved an overall accuracy of 0.98 and a kappa coefficient of 0.94.

Download Full-text

The Value of Sentinel-2 Spectral Bands for the Assessment of Winter Wheat Growth and Development

Remote Sensing ◽

10.3390/rs11172050 ◽

2019 ◽

Vol 11 (17) ◽

pp. 2050 ◽

Cited By ~ 8

Author(s):

Andrew Revill ◽

Anna Florence ◽

Alasdair MacArthur ◽

Stephen Hoad ◽

Robert Rees ◽

...

Keyword(s):

Winter Wheat ◽

Chlorophyll Content ◽

Management Practices ◽

Near Infrared ◽

Spectral Characteristics ◽

Crop Management ◽

Area Index ◽

Spectral Bands ◽

The Mean ◽

Sentinel 2

Leaf Area Index (LAI) and chlorophyll content are strongly related to plant development and productivity. Spatial and temporal estimates of these variables are essential for efficient and precise crop management. The availability of open-access data from the European Space Agency’s (ESA) Sentinel-2 satellite—delivering global coverage with an average 5-day revisit frequency at a spatial resolution of up to 10 metres—could provide estimates of these variables at unprecedented (i.e., sub-field) resolution. Using synthetic data, past research has demonstrated the potential of Sentinel-2 for estimating crop variables. Nonetheless, research involving a robust analysis of the Sentinel-2 bands for supporting agricultural applications is limited. We evaluated the potential of Sentinel-2 data for retrieving winter wheat LAI, leaf chlorophyll content (LCC) and canopy chlorophyll content (CCC). In coordination with destructive and non-destructive ground measurements, we acquired multispectral data from an Unmanned Aerial Vehicle (UAV)-mounted sensor measuring key Sentinel-2 spectral bands (443 to 865 nm). We applied Gaussian processes regression (GPR) machine learning to determine the most informative Sentinel-2 bands for retrieving each of the variables. We further evaluated the GPR model performance when propagating observation uncertainty. When applying the best-performing GPR models without propagating uncertainty, the retrievals had a high agreement with ground measurements—the mean R2 and normalised root-mean-square error (NRMSE) were 0.89 and 8.8%, respectively. When propagating uncertainty, the mean R2 and NRMSE were 0.82 and 11.9%, respectively. When accounting for measurement uncertainty in the estimation of LAI and CCC, the number of most informative Sentinel-2 bands was reduced from four to only two—the red-edge (705 nm) and near-infrared (865 nm) bands. This research demonstrates the value of the Sentinel-2 spectral characteristics for retrieving critical variables that can support more sustainable crop management practices.

Download Full-text

Wildfire Detection Probability of MODIS Fire Products under the Constraint of Environmental Factors: A Study Based on Confirmed Ground Wildfire Records

Remote Sensing ◽

10.3390/rs11243031 ◽

2019 ◽

Vol 11 (24) ◽

pp. 3031 ◽

Cited By ~ 2

Author(s):

Lingxiao Ying ◽

Zehao Shen ◽

Mingzheng Yang ◽

Shilong Piao

Keyword(s):

Environmental Factors ◽

Detection Probability ◽

Ground Level ◽

Omission Error ◽

Limiting Factor ◽

Commission Error ◽

Fire Size ◽

Weather Factors ◽

Local Conditions ◽

Wildfire Occurrence

The Moderate Resolution Imaging Spectroradiometer (MODIS) has been widely used for wildfire occurrence and distribution detecting and fire risk assessments. Compared with its commission error, the omission error of MODIS wildfire detection has been revealed as a much more challenging, unsolved issue, and ground-level environmental factors influencing the detection capacity are also variable. This study compared the multiple MODIS fire products and the records of ground wildfire investigations during December 2002–November 2015 in Yunnan Province, Southwest China, in an attempt to reveal the difference in the spatiotemporal patterns of regional wildfire detected by the two approaches, to estimate the omission error of MODIS fire products based on confirmed ground wildfire records, and to explore how instantaneous and local environmental factors influenced the wildfire detection probability of MODIS. The results indicated that across the province, the total number of wildfire events recorded by MODIS was at least twice as many as that in the ground records, while the wildfire distribution patterns revealed by the two approaches were inconsistent. For the 5145 confirmed ground records, however, only 11.10% of them could be detected using multiple MODIS fire products (i.e., MOD14A1, MYD14A1, and MCD64A1). Opposing trends during the studied period were found between the yearly occurrence of ground-based wildfire records and the corresponding proportion detected by MODIS. Moreover, the wildfire detection proportion by MODIS was 11.36% in forest, 9.58% in shrubs, and 5.56% in grassland, respectively. Random forest modeling suggested that fire size was a primary limiting factor for MODIS fire detecting capacity, where a small fire size could likely result in MODIS omission errors at a threshold of 1 ha, while MODIS had a 50% probability of detecting a wildfire whose size was at least 18 ha. Aside from fire size, the wildfire detection probability of MODIS was also markedly influenced by weather factors, especially the daily relative humidity and the daily wind speed, and the altitude of wildfire occurrence. Considering the environmental factors’ contribution to the omission error in MODIS wildfire detection, we emphasized the importance of attention to the local conditions as well as ground inspection in practical wildfire monitoring and management and global wildfire simulations.

Download Full-text

Mapping of Scotch Broom (Cytisus scoparius) with Landsat Imagery

Weed Technology ◽

10.1614/wt-d-15-00038.1 ◽

2016 ◽

Vol 30 (2) ◽

pp. 539-558 ◽

Cited By ~ 2

Author(s):

David A. Hill ◽

Raj Prasad ◽

Donald G. Leckie

Keyword(s):

Native Species ◽

Landsat Imagery ◽

Vegetation Indices ◽

Commission Error ◽

Invasive Shrub ◽

Use Of Time ◽

Spectral Bands ◽

Scotch Broom ◽

Band Combinations ◽

Reflectance Measurements

Methods were developed and tested for mapping the distribution of Scotch broom, an invasive shrub species expanding its range and disrupting native species and habitats in several parts of the world. During spring, the Scotch broom produces yellow flowers. Landsat imagery during the flower bloom period and during summer was acquired for several years for a study area on Vancouver Island, British Columbia, Canada. Ground-based reflectance measurements plus statistical separability tests were conducted to determine the effectiveness for identifying Scotch broom with Landsat spectral bands, band ratios, vegetation indices, and combinations of bloom and nonbloom imagery. Maximum likelihood classifications of three Scotch broom density classes (dense, ≥ 75% cover; moderate, 25 to 75%; low, 10 to 25%) and other land covers were run with various image and band sets and tested against independent reference sites. Accuracies of classifications using the better band combinations for moderate and dense Scotch broom patches combined were on the order of 80%, with unreliable results for sites of low Scotch broom density. Scotch broom patches less than 0.5 ha were often missed. Some commission error occurred (areas erroneously classified as Scotch broom). Suggested improvements are the use of time series of classifications over multiple years, incorporating knowledge of Scotch broom spread mechanisms or temperature and elevation limitations, and use of higher resolution satellites if the expense warrants it. Despite some limitations, a satellite-based remote sensing approach may be useful for aspects of Scotch broom management.

Download Full-text

Crop Mapping Using Random Forest and Particle Swarm Optimization based on Multi-Temporal Sentinel-2

Remote Sensing ◽

10.3390/rs12091449 ◽

2020 ◽

Vol 12 (9) ◽

pp. 1449

Author(s):

Elahe Akbari ◽

Ali Darvishi Boloorani ◽

Najmeh Neysani Samany ◽

Saeid Hamzeh ◽

Saeid Soufizadeh ◽

...

Keyword(s):

Feature Selection ◽

Particle Swarm Optimization ◽

Importance Measure ◽

Swarm Optimization ◽

Feature Sets ◽

Classification Feature ◽

Spectral Bands ◽

Multi Temporal ◽

Crop Mapping ◽

Sentinel 2

Timely and accurate information on crop mapping and monitoring is necessary for agricultural resources management. Accordingly, the applicability of the proposed classification-feature selection ensemble procedure with different feature sets for crop mapping is investigated. Here, we produced various feature sets including spectral bands, spectral indices, variation of spectral index, texture, and combinations of features to map different types of crops. By using various feature sets and the random forest (RF) classifier, the crop maps were created. In aiming to determine the most relevant and distinctive features, the particle swarm optimization (PSO) and RF-variable importance measure feature selection methods were examined. The classification-feature selection ensemble procedure was adapted to combine the outputs of different feature sets from the better feature selection method using majority votes. Multi-temporal Sentinel-2 data has been used in Ghale-Nou county of Tehran, Iran. The performance of RF was efficient in crop mapping especially by spectral bands and texture in combination with other feature sets. Our results showed that the PSO-based feature selection leads to a more accurate classification than the RF-variable importance measure, in almost all feature sets for all crop types. The RF classifier-PSO ensemble procedure for crop mapping outperformed the RF classifier in each feature set with regard to the class-wise and overall accuracies (OA) (of about 2.7–7.4% increases in OA and 0.48–3.68% (silage maize), 0–1.61% (rice), 2.82–15.43% (alfalfa), and 10.96–41.13% (vegetables) improvement in F-scores for all feature sets). The proposed method could mainly be useful to differentiate between heterogeneous crop fields (e.g., vegetables in this study) due to their more obtained omission/commission errors reduction.

Download Full-text

BEECH TREE DENSITY ESTIMATION USING SENTINEL-2 DATA (CASE STUDY: KHYROUD FOREST)

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xlii-4-w18-891-2019 ◽

2019 ◽

Vol XLII-4/W18 ◽

pp. 891-895

Author(s):

G. Ronoud ◽

A. A. Darvish Sefat ◽

P. Fatehi

Keyword(s):

Density Estimation ◽

Vegetation Indices ◽

Model Performance ◽

Field Measurements ◽

Tree Density ◽

Spectral Bands ◽

Tasseled Cap Transformation ◽

Beech Tree ◽

Remote Sensing Techniques ◽

Sentinel 2

Abstract. Obtaining information about forest attributes is essential for planning, monitoring, and management of forests. Due to the time and cost consuming of Tree Density (TD) using field measurements especially in the vast and remote areas, remote sensing techniques have gained more attention in scientific community. Khyroud forest, a part of Hyrcanian forest of Iran, with a high species biodiversity and growing volume stock plays an important role in carbon storage. The aim of this study was to assess the capability of Sentinel-2 data for estimating the tree density in the Khyroud forest. 65 square sample plots with an area of 2025 m2 were measured. In each sample plot, trees with diameter at the breast height (DBH) higher than 7.5-cm were recorded. The quality of Sentinel-2 data in terms of geometric correction and cloud effect were investigated. Different processing approaches such as vegetation indices and Tasseled Cap transformation on spectral bands in combination with an empirical approach were implemented. Also, some of biophysical variables were computed. To assess the model performance, the data were randomly divided into parts, 70% of sample plots were used for modelling and 30% for validation. The results showed that the SVR algorithm (linear kernel) with a relative RMSE of 23.09% and a R2 of 0.526 gained the highest performance for tree density estimation.

Download Full-text

ASSESSMENTS OF SENTINEL-2 VEGETATION RED-EDGE SPECTRAL BANDS FOR IMPROVING LAND COVER CLASSIFICATION

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-archives-xlii-2-w7-871-2017 ◽

2017 ◽

Vol XLII-2/W7 ◽

pp. 871-874 ◽

Cited By ~ 4

Author(s):

S. Qiu ◽

B. He ◽

C. Yin ◽

Z. Liao

Keyword(s):

Land Cover ◽

Land Cover Classification ◽

Relative Importance ◽

East Texas ◽

Spectral Bands ◽

Red Edge ◽

Sentinel 2A ◽

Land Cover Maps ◽

Confusion Matrices ◽

Sentinel 2

The Multi Spectral Instrument (MSI) onboard Sentinel-2 can record the information in Vegetation Red-Edge (VRE) spectral domains. In this study, the performance of the VRE bands on improving land cover classification was evaluated based on a Sentinel-2A MSI image in East Texas, USA. Two classification scenarios were designed by excluding and including the VRE bands. A Random Forest (RF) classifier was used to generate land cover maps and evaluate the contributions of different spectral bands. The combination of VRE bands increased the overall classification accuracy by 1.40&thinsp;%, which was statistically significant. Both confusion matrices and land cover maps indicated that the most beneficial increase was from vegetation-related land cover types, especially agriculture. Comparison of the relative importance of each band showed that the most beneficial VRE bands were Band 5 and Band 6. These results demonstrated the value of VRE bands for land cover classification.

Download Full-text