Vegetation Index Images Research Articles

Development of a Simple Observation System to Monitor Regional Wind Erosion

Dryland occupies about 46% of the global land surface area (except Antarctica) and is the most vulnerable area to climate change. From the conditions of vegetation and land surface wetness and blown sand phenomena, a simple observation system was developed to monitor regional wind erosion and applied to Khuld of Mongolia, which is sensitive to drought and desertification. The system was composed of instruments that observed blown sand, vegetation amount, land surface wetness, and landscape features related to regional wind erosion. Sixteen blown sand and eight sandstorm events were evaluated from 5 March to 5 June 2023 (i.e., during the Asian dust season in northeast Asia). The normalized difference vegetation index and visible images showed that the vegetation amount was considerably less, and the developed moisture index related to land surface wetness indicated dry conditions. Combining the results of blown sand, these indices, and visible images, land surface conditions during the analysis period were likely to occur with blown sand events.

Global forest resilience change from 2001 to 2022

ABSTRACT Forest resilience, the ecosystem’s capacity to withstand perturbations and retain primary functions and characteristics, is an essential indicator in evaluating the fate of an ecosystem under rapid climate change. Detailed information about forest resilience change, such as number of changes, direction (reduced vs increased), timing (start/end year), or duration of the change, is critical but is often not well demonstrated at the global scale. Here, we applied lag-one autocorrelation (AC1) as a forest resilience indicator on time series MODIS vegetation index images from 2001 to 2022 and employed LandTrendr spectral-temporal segmentation algorithms to track forest resilience change. Our results showed that the resilience of global forests changed nearly 3 times on average, and over 50% of global forests had an overall downward trend from the early 2000s up to recently. However, the current condition is changing for the better globally with approximately 53% of global forests currently showing an increased resilience during the recent change, and average patches of reduced resilience becoming smaller and more scattered. Over half of the forests in arid and temperate domains still show decreased resilience recently, which highlights the need for improved management practices.

Development of an Automated Low-Cost Multispectral Imaging System to Quantify Canopy Size and Pigmentation.

Canopy imaging offers a non-destructive, efficient way to objectively measure canopy size, detect stress symptoms, and assess pigment concentrations. While it is faster and easier than traditional destructive methods, manual image analysis, including segmentation and evaluation, can be time-consuming. To make imaging more widely accessible, it's essential to reduce the cost of imaging systems and automate the analysis process. We developed a low-cost imaging system with automated analysis using an embedded microcomputer equipped with a monochrome camera and a filter for a total hardware cost of ~USD 500. Our imaging system takes images under blue, green, red, and infrared light, as well as chlorophyll fluorescence. The system uses a Python-based program to collect and analyze images automatically. The multi-spectral imaging system separates plants from the background using a chlorophyll fluorescence image, which is also used to quantify canopy size. The system then generates normalized difference vegetation index (NDVI, "greenness") images and histograms, providing quantitative, spatially resolved information. We verified that these indices correlate with leaf chlorophyll content and can easily add other indices by installing light sources with the desired spectrums. The low cost of the system can make this imaging technology widely available.

Focal-TSMP: deep learning for vegetation health prediction and agricultural drought assessment from a regional climate simulation

Abstract. Satellite-derived agricultural drought indices can provide a complementary perspective of terrestrial vegetation trends. In addition, their integration for drought assessments under future climates is beneficial for providing more comprehensive assessments. However, satellite-derived drought indices are only available for the Earth observation era. In this study, we aim to improve the agricultural drought assessments under future climate change by applying deep learning (DL) to predict satellite-derived vegetation indices from a regional climate simulation. The simulation is produced by the Terrestrial Systems Modeling Platform (TSMP) and performed in a free evolution mode over Europe. TSMP simulations incorporate variables from underground to the top of the atmosphere (ground-to-atmosphere; G2A) and are widely used for research studies related to water cycle and climate change. We leverage these simulations for long-term forecasting and DL to map the forecast variables into normalized difference vegetation index (NDVI) and brightness temperature (BT) images that are not part of the simulation model. These predicted images are then used to derive different vegetation and agricultural drought indices, namely NDVI anomaly, BT anomaly, vegetation condition index (VCI), thermal condition index (TCI), and vegetation health index (VHI). The developed DL model could be integrated with data assimilation and used for downstream tasks, i.e., for estimating the NDVI and BT for periods where no satellite data are available and for modeling the impact of extreme events on vegetation responses with different climate change scenarios. Moreover, our study could be used as a complementary evaluation framework for TSMP-based climate change simulations. To ensure reliability and to assess the model’s applicability to different seasons and regions, we provide an analysis of model biases and uncertainties across different regions over the pan-European domain. We further provide an analysis about the contribution of the input variables from the TSMP model components to ensure a better understanding of the model prediction. A comprehensive evaluation of the long-term TSMP simulation using reference remote sensing data showed sufficiently good agreements between the model predictions and observations. While model performance varies on the test set between different climate regions, it achieves a mean absolute error (MAE) of 0.027 and 1.90 K with coefficient of determination (R2) scores of 0.88 and 0.92 for the NDVI and BT, respectively, at 0.11° resolution for sub-seasonal predictions. In summary, we demonstrate the feasibility of using DL on a TSMP simulation to synthesize NDVI and BT satellite images, which can be used for agricultural drought forecasting. Our implementation is publicly available at the project page (https://hakamshams.github.io/Focal-TSMP, last access: 4 April 2024).

Soil organic carbon stocks in native forest of Argentina: a useful surrogate for mitigation and conservation planning under climate variability

BackgroundThe nationally determined contribution (NDC) presented by Argentina within the framework of the Paris Agreement is aligned with the decisions made in the context of the United Nations Framework Convention on Climate Change (UNFCCC) on the reduction of emissions derived from deforestation and forest degradation, as well as forest carbon conservation (REDD+). In addition, climate change constitutes one of the greatest threats to forest biodiversity and ecosystem services. However, the soil organic carbon (SOC) stocks of native forests have not been incorporated into the Forest Reference Emission Levels calculations and for conservation planning under climate variability due to a lack of information. The objectives of this study were: (i) to model SOC stocks to 30 cm of native forests at a national scale using climatic, topographic and vegetation as predictor variables, and (ii) to relate SOC stocks with spatial–temporal remotely sensed indices to determine biodiversity conservation concerns due to threats from high inter-annual climate variability.MethodsWe used 1040 forest soil samples (0–30 cm) to generate spatially explicit estimates of SOC native forests in Argentina at a spatial resolution of approximately 200 m. We selected 52 potential predictive environmental covariates, which represent key factors for the spatial distribution of SOC. All covariate maps were uploaded to the Google Earth Engine cloud-based computing platform for subsequent modelling. To determine the biodiversity threats from high inter-annual climate variability, we employed the spatial–temporal satellite-derived indices based on Enhanced Vegetation Index (EVI) and land surface temperature (LST) images from Landsat imagery.ResultsSOC model (0–30 cm depth) prediction accounted for 69% of the variation of this soil property across the whole native forest coverage in Argentina. Total mean SOC stock reached 2.81 Pg C (2.71–2.84 Pg C with a probability of 90%) for a total area of 460,790 km2, where Chaco forests represented 58.4% of total SOC stored, followed by Andean Patagonian forests (16.7%) and Espinal forests (10.0%). SOC stock model was fitted as a function of regional climate, which greatly influenced forest ecosystems, including precipitation (annual mean precipitation and precipitation of warmest quarter) and temperature (day land surface temperature, seasonality, maximum temperature of warmest month, month of maximum temperature, night land surface temperature, and monthly minimum temperature). Biodiversity was influenced by the SOC levels and the forest regions.ConclusionsIn the framework of the Kyoto Protocol and REDD+, information derived in the present work from the estimate of SOC in native forests can be incorporated into the annual National Inventory Report of Argentina to assist forest management proposals. It also gives insight into how native forests can be more resilient to reduce the impact of biodiversity loss.

The Use of Sentinel-2 Vegetation Indices Imagery in Detecting the Effect of Plant Distance to the Productivity of Corn Crops

There are several variations in planting distance used for corn plants, including 70×40 cm, 70×20 cm, and irregular. This affects the productivity of corn plants because of competition between corn plants. Satellite imagery, such as Sentinel-2 imagery, can be used to observe this effect. This research aimed to determine the effectiveness of Sentinel-2 satellite imagery in detecting the effect of plant distance on corn plant productivity. This research used Modified Red-edge Simple Ratio (mRE-SR) Vegetation Index imagery from Sentinel-2 imagery. The three plant distances used were 70×40 cm, 70×20 cm, and irregular. The images used corresponded to the corn planting times at research location, i.e. 10 and 20 December 2022, 14 January 2023, 28 February 2023 and 15 March 2023. Linear regression analysis was carried out to test the correlation between corn plant productivity of each plant distance and the mRE-SR Vegetation Index image. The results of this research indicated that the effect of plant distance on corn plant productivity can be studied using mRE-SR Sentinel-2 imagery.

Changes in the relationship between Index of Concentration at the Extremes and U.S. urban greenspace: a longitudinal analysis from 2001–2019

Urban greenspace is associated with a wide variety of human health benefits, from improved mental health to reduced violent crime and infant birth outcomes. This study investigates long-term trends in the distribution of greenspace across census tracts in the 260 U.S. cities with populations greater than 100,000 residents based on economic privilege and deprivation. Satellite-acquired Normalized Difference Vegetation Index (NDVI) images were used as a measure of greenspace, and the Index of Concentration at the Extremes (ICE) was used as a measure spatial social polarizations of deprived and priveliged populations. From 2001 to 2019, the most privileged, highest ICE quintile census tracts had both the highest mean NDVI and experienced a significantly greater increase in greenspace than the least privileged tracts, indicating that the disparity in greenspace access is widening. Public greenspace initiatives need to focus on programs that are more equitable across the spectrum of economic strata to reduce urban health disparities and address concerns of environmental justice.

Remote sensing of evapotranspiration for irrigated crops at Yuma, Arizona, USA

A satellite-based vegetation index model that tracks daily crop growth and evapotranspiration (ETc) is developed, tested, and validated over irrigated farms in Yuma irrigation districts of Arizona and California. Model inputs are remotely sensed normalized difference vegetation index (NDVI) images, crop type maps, and local weather. The utility and novelty of the model is a more accurate assessment of ETc than currently provided by the US Bureau of Reclamation’s evapotranspiration modeling system. The model analyzes NDVI time series data from the European Space Agency’s Sentinel-2 satellites using the Google Earth Engine, constructs FAO-56 style crop growth stages from NDVI, and then estimates daily ETc using pre-defined crop coefficients (Kc) and grass reference evapotranspiration (ETos). Four crops were selected to test and evaluate model performance: short-season broccoli, mid-season cotton and wheat, and perennial alfalfa. Comparison of model results showed that Reclamation reports overestimate alfalfa and wheat ETc by 21–25%, cotton ETc by 6%, and underestimate broccoli ETc by 21%. Variability resolved by the model ranged 6–18% of median total ETc. Comparison of model results with those obtained from 13 eddy covariance sites showed validation discrepancies ranging 1–14%: average total actual ETc differences were 12, − 14, 78, and 87 mm/season, respectively, for alfalfa, broccoli, cotton, and wheat. The wide availability of Sentinel-2 data, collected every 5 days or less, and the rapid processing via Google Earth Engine make the vegetation index model implementation fast and practical. Its accuracy and ability to resolve ETc for every field would benefit the Reclamation water accounting system and provide valuable consumptive water use data for any Colorado River stakeholder.

Design of a Multilayer Perceptron Network Based on the Normalized Histogram to Develop Yields Predictive Model

This article presents a multilayer perceptron (MLP) for patio-temporal remote sensing analysis of satellite image time series for a yield prediction model. This paper illustrates the model for crop yield prediction from the spatial cumulation of normalized difference vegetation index (NDVI) image time series at the province level. The methodological framework comprises the transformation of each NDVI image into a histogram to ensure no loss of information in the mapping of high- dimensional/unstructured NDVI images into pixels consideration, to be then used in training the MLP. The research work also includes an analysis of several activation functions for the hidden layer and testing their consequences on accuracy, including Radial-Basis (RadBas), Logarithmic-Sigmoid (LogSig), Hyperbolic-Tangent Sigmoid (TanSig), which depend mainly on exponential functions and limit the amplitude of the output. The proposed approach was utilized to predict the winter crop yield in Diwaniyah-Iraq province, one of the main agriculture regions, whose economy considerably depends on crop production. It can also be extended to other crops and other regions in Iraq. The proposed methodology showed the ability to predict the crop yield around seven-nine weeks before harvest. It also, outperformed the performance of traditional approaches by transforming the input into a more convenient form that reflects more useful information. The results show that the proposed model provides efficient accuracy (determination coefficient R2 > 0.85 and error level <0.24).

Monitoring vegetation dynamics using multi-temporal Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) images of Tamil Nadu

Vegetation indices serve as an essential tool in monitoring variations in vegetation. The vegetation indices used often, viz., normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) were computed from MODIS vegetation index products. The present study aimed to monitor vegetation's seasonal dynamics by using time series NDVI and EVI indices in Tamil Nadu from 2011 to 2021. Two products characterize the global range of vegetation states and processes more effectively. The data sources were processed and the values of NDVI and EVI were extracted using ArcGIS software. There was a significant difference in vegetation intensity and status of vegetation over time, with NDVI having a larger value than EVI, indicating that biomass intensity varies over time in Tamil Nadu. Among the land cover classes, the deciduous forest showed the highest mean values for NDVI (0.83) and EVI (0.38), followed by cropland mean values of NDVI (0.71) and EVI (0.31) and the lowest NDVI (0.68) and EVI (0.29) was recorded in the scrubland. The study demonstrated that vegetation indices extracted from MODIS offered valuable information on vegetation status and condition at a short temporal time period.

Use of Plant Health Level Based on Random Forest Algorithm for Agricultural Drone Target Points

Chemical residues from the use of pesticides in agriculture can impact human health through environmental and food pollution. To lessen the negative effects of excessive pesticide use, pesticides must be applied to plants by dose. The dose of pesticide application can be based on a plant health level, which is the result of drone Normalized Difference Vegetation Index (NDVI) image analysis. Drones can also be used for spraying pesticides. Analysis of plant health levels was carried out using the Random Forest (RF) algorithm. The results of the classification plant health levels will be used to design spray drone flight routes. The objective of this research is to classify plant health levels of rice based on NDVI imagery using the RF algorithm and to compile a database of spray drone target points. The results of this study indicate that the classification of plant health levels using the RF algorithm produces an accuracy value of 98% and a Kappa value of 0.96. As a result, the model developed and the algorithm employed is quite effective at classifying the level of plant health. Furthermore, spray drone target points based on plant health levels can be generated. Optimally the spray distance between rows is 2 m.

DROUGHT MONITORING FROM 2001–2019 IN NORTHEAST THAILAND USING MODIS NDVI IMAGE TIME SERIES AND Savitzky-Golay APPROACH

Abstract. Drought directly threatens food security and livelihoods, thereby increasing socioeconomic risks and remains a challenge for natural resource management, particularly in frequently affected regions. Earth observation (EO) satellites provide extensive spectral and temporal data for long-term drought monitoring. This study monitored droughts in Northeast Thailand from 2001 to 2019 using the MODIS normalised difference vegetation index (NDVI) image time series. The Savitzky-Golay (S-G) method was used to remove noise and fill gaps in the image datasets. Optimal indicators as the vegetation condition index (VCI) and the standard vegetation index (SVI) were used to monitor drought distribution patterns over the previous 19 years. S-G filtering effectively reduced the impact of undetected clouds and water vapour, while VCI had the highest accuracy coefficient of determination (R2) for rainfall data at 0.85. Long-term droughts occurred frequently in 2005, 2004, 2007, and 2001 with the northern and central regions most severely affected. Severe drought primarily impacted agricultural land, forest and miscellaneous areas. Inter-annual drought variability for one and three time steps was clearly demonstrated in May and April to June from 2001 to 2019. Overall, the VCI provided a high level of satisfaction for drought monitoring in this region and clearly displayed the spatial distribution of long-term drought regions. Our findings provide a valuable resource for drought mitigation planning and warning systems.

A simple method for the estimation of minimum and maximum air temperature monthly mean maps using MODIS images in the region of Murcia, Spain

Air temperature records are acquired by networks of weather stations which may be several kilometres apart. In complex topographies the representativeness of a meteorological station may be diminished in relation to a flatter valley, and the nearest station may have no relation to a place located near it. The present study shows a simple method to estimate the spatial distribution of minimum and maximum air temperatures from MODIS land surface temperature (LST) and normalized difference vegetation index (NDVI) images. Indeed, there is a strong correlation between MODIS day and night LST products and air temperature records from meteorological stations, which is obtained by using geographically weighted regression equations, and reliable results are found. Then, the results allow to spatially interpolate the coefficients of the local regressions using altitude and NDVI as descriptor variables, to obtain maps of the whole region for minimum and maximum air temperature. Most of the meteorological stations show air temperature estimates that do not have significant differences compared to the measured values. The results showed that the regression coefficients for the selected locations are strong for the correlations between minimum temperature with LSTnight (R2 = 0.69–0.82) and maximum temperature with LSTday (R2 = 0.70–0.87) at the 47 stations. The root mean square errors (RMSE) of the statistical models are 1.0 °C and 0.8 °C for night and daytime temperatures, respectively. Furthermore, the association between each pair of data is significant at the 95% level (p<0.01).

Forest Land Resource Information Acquisition with Sentinel-2 Image Utilizing Support Vector Machine, K-Nearest Neighbor, Random Forest, Decision Trees and Multi-Layer Perceptron

Forestry work involves scientific management and the effective utilization of forest land resources, and finding economical, efficient and accurate acquisition methods for forest land resource information. In previous land-use classification research, machine learning algorithms have achieved good results, and Sentinel public data have been used in various remote sensing applications. However, there is a paucity of research using these data to evaluate the performance of machine learning algorithms in the extracting of complex forest land resource information. Using the Sentinel-2 satellite multispectral image data, the spectral reflectance, vegetation index characteristics and image texture characteristics of different forest land resources in the study area were calculated and compared. Then, based on three groups of features, the performances of the Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Random Forest (RF), decision trees (DT) and Multi-layer Perceptron (MLP) were examined and compared to identify and classify forest land resource types. The research indicates the following: (1) The SVM algorithm achieved the highest OA (95.8%). The average accuracy of the SVM algorithm was much higher than other algorithms (SVM 88.3%, KNN 87.5, RF 85.3%, MLP 85.00% and DT 77.5%). (2) The classification accuracies of each algorithm for coniferous forests were relatively high, and the recognition accuracy was above 95%, whereas the classification accuracies of the other categories varied greatly. (3) Adding texture features can improve the accuracy of the five algorithms. This study reports new references for the qualitative methods of forest land resource distribution. It has also produced more efficient and accurate acquisitions of forest land resource information, scientific management and effective use of forest land resources.

Development of a Monitoring Method Using UAVs That Can Detect the Occurrence of Bark Stripping by Deer

The occurrence of bark stripping associated with increased deer densities can severely damage forests. Identifying trends in bark stripping is crucial for forest management, but such data are often difficult to obtain through field surveys. Therefore, this study aimed to develop an efficient monitoring method using unmanned aerial vehicles (UAVs) that can detect the occurrence of bark stripping and enable long-term monitoring. The area around the Ochiai Pass in Higashi-Iya Ochiai, Miyoshi City, Tokushima Prefecture, Japan, was selected as the study area for the survey of Abies homolepis, which was found to be significantly bark-stripped by deer in the field. The location and root diameter of A. homolepis were measured, and the percentages of bark stripping and tree growth were visually determined. Simultaneously, normalized difference vegetation index (NDVI) and visible light orthomosaic images were produced using a UAV. A canopy polygon of A. homolepis was created, and the average value of the NDVI within the polygon was calculated. Where the bark stripping rate at the root edge was greater than 75%, the number of “partially dead” and “dead” trees increased significantly, indicating that bark stripping by deer was the primary cause of the death of A. homolepis in Ochiai Pass. In addition, the mean value of the NDVI was significantly lower, with a bark stripping rate of 75% or higher, indicating that the NDVI of the canopy of A. homolepis can be used to estimate individuals with a high bark stripping rate at the root tips, that is, those with a high probability of mortality. Furthermore, by extrapolating the results of the tree-by-tree survey to the nontarget A. homolepis, we detected 46 (8%) A. homolepis with an average NDVI value of 0.8 or less (i.e., those with a bark stripping ratio of 75% or higher and a high probability of mortality). Therefore, the utilization of remote sensing technology via UAVs, as demonstrated in this study, proves to be a potent means for monitoring the incidence of bark stripping.

The Impact of the Type and Spatial Resolution of a Source Image on the Effectiveness of Texture Analysis

This paper presents a study on the effectiveness of texture analysis of remote sensing imagery depending on the type and spatial resolution of the source image. The study used the following image types: near-infrared band, red band, first principal component, second principal component and normalized difference vegetation index images of pixel size from 2 m to 30 m, generated from a multispectral WorldView-2 image. The study evaluated the separability of the selected pairs of the following land cover classes: bare soil, low vegetation, coniferous forest, deciduous forest, water reservoirs, built-up areas. The tool used for texture analysis was granulometric analysis based on morphological operations—one of less popular methods which, however, as demonstrated by previous studies, shows high effectiveness in separating classes of different texture. The conducted study enabled researchers to evaluate the significance of image type and resolution for visibility of texture in the image and the possibility of using texture to differentiate between classes. The obtained results showed that there is no single, universal combination of conditions of texture analysis, which would be the best from the point of view of all classes. For most of the analyzed pairs of classes, the best results were obtained for the highest spatial resolution of the image (2–3 m), but the class of built-up areas stands out in this comparison—the best distinction was obtained with the average spatial resolution (10–15 m). Research has also shown that there is no single type of image that is universally the best basis for texture analysis. While for the majority of classes the image of the first principal component was the best, for the class of built-up areas it was the image of the red channel.

Immunity to poliovirus in Afghanistan: A household sampling method for serological assessment based on geographical information systems.

Afghanistan continues to experience challenges affecting polio eradication. Mass polio vaccination campaigns, which aim to protect children under the age of 5, are a key eradication strategy. To date, the polio program in Afghanistan has only employed facility-based seroprevalence surveys, which can be subject to sampling bias. We describe the feasibility in implementing a cross-sectional household poliovirus seroprevalence survey based on geographical information systems (GIS) in three districts. Digital maps with randomly selected predetermined starting points were provided to teams, with a total target of 1,632 households. Teams were instructed to navigate to predetermined starting points and enrol the closest household within 60 m. To assess effectiveness of these methods, we calculated percentages for total households enrolled with valid geocoordinates collected within the designated boundary, and whether the Euclidean distance of households were within 60 m of a predetermined starting point. A normalized difference vegetation index (NDVI) image ratio was conducted to further investigate variability in team performances. The study enrolled a total of 78% of the target sample with 52% of all households within 60 m of a pre-selected point and 79% within the designated cluster boundary. Success varied considerably between the four target areas ranging from 42% enrolment of the target sample in one place to 90% enrolment of the target sample in another. Interviews with the field teams revealed that differences in security status and amount of non-residential land cover were key barriers to higher enrolment rates. Our findings indicate household poliovirus seroprevalence surveys using GIS-based sampling can be effectively implemented in polio endemic countries to capture representative samples. We also proposed ways to achieve higher success rates if these methods are to be used in the future, particularly in areas with concerns of insecurity or spatially dispersed residential units.

Study on the spatiotemporal dynamic of ground-level ozone concentrations on multiple scales across China during the blue sky protection campaign

Surface ozone (O3), one of the harmful air pollutants, generated significantly negative effects on human health and plants. Existing O3 datasets with coarse spatiotemporal resolution and limited coverage, and the uncertainties of O3 influential factors seriously restrain related epidemiology and air pollution studies. To tackle above issues, we proposed a novel scheme to estimate daily O3 concentrations on a fine grid scale (1 km × 1 km) from 2018 to 2020 across China based on machine learning methods using hourly observed ground-level pollutant concentrations data, meteorological data, satellite data, and auxiliary data including digital elevation model (DEM), land use data (LUD), normalized difference vegetation index (NDVI), population (POP), and nighttime light images (NTL), and to identify the difference of influential factors of O3 on diverse urbanization and topography conditions. Some findings were achieved. The correlation coefficients (R2) between O3 concentrations and surface net solar radiation (SNSR), boundary layer height (BLH), 2 m temperature (T2M), 10 m v-component (MVW), and NDVI were 0.80, 0.40, 0.35, 0.30, and 0.20, respectively. The random forest (RF) demonstrated the highest validation R2 (0.86) and lowest validation RMSE (13.74 μg/m3) in estimating O3 concentrations, followed by support vector machine (SVM) (R2 = 0.75, RMSE = 18.39 μg/m3), backpropagation neural network (BP) (R2 = 0.74, RMSE = 19.26 μg/m3), and multiple linear regression (MLR) (R2 = 0.52, RMSE = 25.99 μg/m3). Our China High-Resolution O3 Dataset (CHROD) exhibited an acceptable accuracy at different spatial–temporal scales. Additionally, O3 concentrations showed decreasing trend and represented obviously spatiotemporal heterogeneity across China from 2018 to 2020. Overall, O3 was mainly affected by human activities in higher urbanization regions, while O3 was mainly controlled by meteorological factors, vegetation coverage, and elevation in lower urbanization regions. The scheme of this study is useful and valuable in understanding the mechanism of O3 formation and improving the quality of the O3 dataset.

Detection of fusarium head blight in wheat under field conditions using a hyperspectral camera and machine learning

Fusarium head blight (FHB) is among the most devastating fungal diseases in cereal crops, reducing yield, and affecting human and livestock health through the production of mycotoxin. Despite application of fungicides, complete eradication of disease is virtually impossible in the field. There is a need for a disease detection technology during late growing stage for estimation of yield affected with FHB and for potential selective harvesting. Most published studies have focused on FHB detection during the milk growth stage using hyperspectral cameras. This preliminary study attempted to fill the knowledge gap by detecting FHB at the ripening stage. A spectral library of healthy and infected ears was collected with a hyperspectral camera in the visible and near-infrared region, over the canopy of eight different wheat varieties. The ears were segmented from the background using a simple linear iterative clustering (SLIC) superpixel algorithm on the normalized difference vegetation index (NDVI) images. Three different machine learning methods, namely, support vector machine (SVM), artificial neural network (ANN), and logistic regression (LR), were utilized for classification. To visualize the FHB distribution in the hypercube, the best performing model was applied for predicting the infected ears in the canopy images. The percentage area coverage of FHB for each hypercube was estimated. Results showed that the SVM algorithm produced the best classification accuracy (CA) of 95.6 % in the test set, followed successively by ANN and LR with CA values of 82.9 and 82.5 %, respectively. Interestingly, the preliminary study shows significant differences in spectral reflectance according to the variety of different resistance levels. The study also proves the feasibility of FHB detection using the developed prediction model during late growth stage with the potential of yield loss estimation before harvest.

Poyang Lake Wetland Classification Using Time-Series ENVISAT ASAR Data and Beijing-1 Imagery

Beijing-1 and ENVISAT ASAR images were used to classify wetland aquatic macrophytes in terms of their plant functional types (PFTs) over the Poyang Lake region, China. Speckle noise filtering, systematic sensor calibration within the same polarization or between different polarizations, and accurate geo-registration were applied to the time-series SAR data. As a result, time-series backscattering data, which is described as permittivity curves in this paper, were obtained. In addition, time-series indices, described as phenological curves, were derived from Beijing-1 time-series images in the classification experiment. Based on these two curves, a rule-based classification strategy was developed to extract wetland information from the combined SAR and optical data. In the rule-based wetland classification method, DEM data, submersion time index, temporal Beijing-1 images, time-series normalized difference vegetation index (TSNDVI) images, principal component analysis (PCA), and temporal ratio of ASAR time-series images were used. In addition, a decision tree-based method was used to map the wetlands. Conclusions include the following: (1) after the preprocessing of ASAR data, it was possible to satisfactorily separate different aquatic plant functional types; (2) hydrophytes from different PFTs exhibited distinct phenological, structural, moisture, and roughness characteristics due to the impact of the annual inundation of Poyang Lake wetland; and (3) more accurate results were obtained with the rule-based method than the decision tree (DT) method. Producer’s and user’s accuracy calculated from test samples in the classification results indicate that the DT method can potentially be used for mapping aquatic PFTs, with overall producer’s accuracy exceeding 80% and higher user’s accuracy for aquatic bed wetland PFTs. A comparison of producer’s and user’s accuracy from the rule-based classification increased from 3 to 12% and 7 to 26%, respectively, for different aquatic PFTs.