Normalized Difference Red Edge Research Articles

Examining the Percent Canopy Cover and Health of Winter Wheat in No-Till and Conventional Tillage Plots Using a Drone

The percent canopy cover and health of winter wheat are important crop performance indicators. Thus, understanding how tillage management practices affect these indicators is beneficial for improving crop performance and consequently yield. The availability of high-resolution drone data with spectral characteristics provides an opportunity to examine the percent canopy cover and health of winter wheat in different tillage systems. This is because the use of drones provides real-time high spatial resolution and temporal images to effectively monitor winter wheat conditions throughout the growing season. Nonetheless, very limited studies have utilized drone data for assessing the percent canopy cover and health conditions of winter wheat for different tillage practices. This study aimed to examine the percent canopy cover and health of winter wheat in no-till and conventional tillage plots using a drone. We used the mean Normalized Difference Vegetation Index (NDVI) ± Standard Deviation (SD) (0.89 ± 0.04) of winter wheat for the growth stages of tillering, jointing, and boot/heading to generate the percent wheat canopy cover. The Normalized Difference Red-Edge (NDRE) produced for winter wheat at the middle and late growth stages was used as a proxy for wheat health condition. We found that the mean percentage canopy cover of winter wheat was about 4% higher in no-till compared to conventional tillage plots in most of the growing season. The mean NDRE ± standard error (SE) of winter wheat was about 0.44 ± 0.01 and 0.43 ± 0.01 for no-till and conventional tillage plots, respectively, during the mid- and late growth stages. There was no significant difference in either the percent canopy cover or health of winter wheat between no-till and conventional tillage plots. The results generated in this study could be used to support farmers’ decision-making process regarding tillage practices and wheat crop performance.

Precision Estimation of Crop Coefficient for Maize Cultivation Using High-Resolution Satellite Imagery to Enhance Evapotranspiration Assessment in Agriculture.

The estimation of crop evapotranspiration (ETc) is crucial for irrigation water management, especially in arid regions. This can be particularly relevant in the Po Valley (Italy), where arable lands suffer from drought damages on an annual basis, causing drastic crop yield losses. This study presents a novel approach for vegetation-based estimation of crop evapotranspiration (ETc) for maize. Three years of high-resolution multispectral satellite (Sentinel-2)-based Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Normalized Difference Red Edge Index (NDRE), and Leaf Area Index (LAI) time series data were used to derive crop coefficients of maize in nine plots at the Acqua Campus experimental farm of Irrigation Consortium for the Emilia Romagna Canal (CER), Italy. Since certain vegetation indices (VIs) (such as NDVI) have an exponential nature compared to the other indices, both linear and power regression models were evaluated to estimate the crop coefficient (Kc). In the context of linear regression, the correlations between Food and Agriculture Organization (FAO)-based Kc and NDWI, NDRE, NDVI, and LAI-based Kc were 0.833, 0.870, 0.886, and 0.771, respectively. Strong correlation values in the case of power regression (NDWI: 0.876, NDRE: 0.872, NDVI: 0.888, LAI: 0.746) indicated an alternative approach to provide crop coefficients for the vegetation period. The VI-based ETc values were calculated using reference evapotranspiration (ET0) and VI-based Kc. The weather station data of CER were used to calculate ET0 based on Penman-Monteith estimation. Out of the Vis, NDWI and NDVI-based ETc performed the best both in the cases of linear (NDWI RMSE: 0.43 ± 0.12; NDVI RMSE: 0.43 ± 0.095) and power (NDWI RMSE: 0.44 ± 0.116; NDVI RMSE: 0.44 ± 0.103) approaches. The findings affirm the efficacy of the developed methodology in accurately assessing the evapotranspiration rate. Consequently, it offers a more refined temporal estimation of water requirements for maize cultivation in the region.

Integrating plant morphological traits with remote-sensed multispectral imageries for accurate corn grain yield prediction.

Sustainable crop production requires adequate and efficient management practices to reduce the negative environmental impacts of excessive nitrogen (N) fertilization. Remote sensing has gained traction as a low-cost and time-efficient tool for monitoring and managing cropping systems. In this study, vegetation indices (VIs) obtained from an unmanned aerial vehicle (UAV) were used to detect corn (Zea mays L.) response to varying N rates (ranging from 0 to 208 kg N ha-1) and fertilizer application methods (liquid urea ammonium nitrate (UAN), urea side-dressing and slow-release fertilizer). Four VIs were evaluated at three different growth stages of corn (V6, R3, and physiological maturity) along with morphological traits including plant height and leaf chlorophyll content (SPAD) to determine their predictive capability for corn yield. Our results show no differences in grain yield (average 13.2 Mg ha-1) between furrow-applied slow-release fertilizer at ≥156 kg N ha-1 and 208 kg N ha-1 side-dressed urea. Early season remote-sensed VIs and morphological data collected at V6 were least effective for grain yield prediction. Moreover, multivariate grain yield prediction was more accurate than univariate. Late-season measurements at the R3 and mature growth stages using a combination of normalized difference vegetation index (NDVI) and green normalized difference vegetation index (GNDVI) in a multilinear regression model showed effective prediction for corn yield. Additionally, a combination of NDVI and normalized difference red edge index (NDRE) in a multi-exponential regression model also demonstrated good prediction capabilities.

A chlorophyll-constrained semi-empirical model for estimating leaf area index using a red-edge vegetation index

The parameter of leaf area index (LAI) provides valuable information about plant growth. Most remote sensing methods for estimating LAI are based on the vegetation indices (VIs) in the visible and near-infrared (VNIR) spectral regions. Recently, these VIs calculated with one or more bands in the red-edge region become popular in estimating LAI. Since the reflectance in the red-edge spectral region is sensitive to both LAI and leaf chlorophyll content (LCC), most red-edge-based VIs (such as normalized difference red edge index, NDRE) should be used to estimate the canopy chlorophyll content (the product of LAI and LCC) in theory. When applying NDRE to estimating LAI alone, their relationship is affected by the variation in LCC. To mitigate the LCC effect on LAI estimation using NDRE, this study proposed a chlorophyll-constrained semi-empirical model (CSE). The coefficients of the CSE model were determined from the simulated dataset based on radiative transfer models, and hence were not dependent on measured data. The performance of the CSE model on LAI estimation was evaluated using simulated and measured datasets, including a crop dataset collected in China and a multi-ecosystem dataset obtained from the National Ecological Observatory Network (NEON). The results showed that the root mean square error (RMSE) obtained by the CSE model ranged from 0.58 to 0.72 m2/m2 for the simulated and measured datasets. In particular, the CSE model also exhibited wide generality across multiple vegetation types for the NEON dataset, with the RMSE in the range of 0.33–0.86 m2/m2. In comparison, without considering the LCC effect, the model would lead to clear underestimation or overestimation of LAI (RMSE = 1.26–2.11 m2/m2). These findings improved our understanding of the LCC effects on universal LAI predictive models and would facilitate the accurate mapping of LAI over large regions for monitoring plant growth status.

Crop Type Classification using Sentinel 2A-Derived Normalized Difference Red Edge Index (NDRE) and Machine Learning Approach

Abstract Satellite remote sensing (RS) enables the extraction of vital information on land cover and crop type. Land cover and crop type classification using RS data and machine learning (ML) techniques have garnered considerable attention in the scientific community. This study aimed to enhance the field of remote sensing research using high-resolution satellite imagery and the ML approach. To reach this goal, machine learning techniques were used to demonstrate whether they could accurately classify different types of crops in farming areas using the Normalized Difference Red Edge Index (NDRE) from Sentinel 2A. Five ML classifiers, namely Support Vector Machines (SVM), Random Forest (RF), Decision Tree (DT), K-Nearest Neighbors (KNN), and Multi-Layer Perceptron (MLP) were implemented using Python programming on Google Colaboratory. The target crop classes included cereals, fallow, forage, fruits, grassland-pasture, legumes, maize, sugar beet, onion-garlic, sunflower, and watermelon-melon. The classification models exhibited strong performance, as evidenced by their robust overall accuracy (OA). The RF model outperformed, with an OA rate of 95% and a Kappa score of 92%. It was followed by DT, KNN, SVM, and MLP with an OA of 88%, 87%, 85%, and 82%, respectively. This study confirmed the potential enhancement of the application of high spatial resolution satellite RS data and ML for crop type classification, especially in regions that have received less attention in previous studies.

Assessing vegetation indices and productivity across nitrogen gradients: a comparative study under transplanted and direct-seeded rice

Nitrogen responses vary under diverse agronomic management practices, influencing vegetation indices (VIs) and productivity across different ecological conditions. However, the proper quantification of these responses under various crop establishment methods with varied nitrogen levels is rarely studied. Therefore, a field experiment was conducted to investigate the impact of varying nitrogen levels on VIs, growth parameters, yield attributes, yield, and economic aspects of transplanted rice (TR) and direct-seeded rice (DSR). The experiment was conducted in the randomized block design consisted seven N levels, which included 0% recommended dose of nitrogen (RDN) or no nitrogen (N0), 33.33% RDN (N1), 66.66% RDN (N2), 100% RDN (N3), 133.33% RDN (N4), 166.66% RDN (N5) and 200% RDN (N6), and replicated thrice. The plots with higher N levels demonstrated increased values of VIs and treatment N3 (120 kg N ha−1), N4 (160 kg N ha−1), N5 (200 kg N ha−1), and N6 (240 kg N ha−1) showed no statistically significant differences in NDVI (normalized difference vegetation index), RVI (ratio vegetation index), NDRE (normalized difference red edge), and GNDVI (green normalized difference vegetation index) values across the various growth stages of rice. The application of treatment N4 resulted in the highest number of panicles m−2 (348.2 in TR, 376.8 in DSR), filled grains panicle−1 (74.55 in TR, 62.43 in DSR), and a 1,000-grain weight of 26.92 g in TR and 26.76 g in DSR. The maximum yield (4.89 t ha−1) was obtained in transplanted rice at treatment N4 and, 8.15% yield reduction was noted in DSR for the same treatment, which was statistically equivalent to N3, but significantly superior to other N levels. Conversely, in DSR with RDN (120 kg N ha−1), the cost–benefit ratio surpassed that of TR by 16.96%, signifying DSR’s adaptability for more profitable rice cultivation in the region. This research provides valuable insights into optimizing nitrogen management practices for TR and DSR, thereby enhancing rice crop performance and economic returns.

Capturing accurate kelp canopy extent: integrating tides, currents, and species-level morphology in kelp remote sensing

Surface-canopy forming kelps (Macrocystis pyrifera and Nereocystis luetkeana) can be monitored along the Northeast Pacific coast using remote sensing. These kelp canopies can be submerged by tides and currents, making it difficult to accurately determine their extent with remote sensing techniques. Further, both species have morphologically distinct canopies, each made up of structures with differing buoyancies, and it is not well understood whether the differing buoyancies between these species’ canopies affects their detectability with remote sensing technologies. Here, we collected in situ above-water spectral signatures for the surface-canopies of Nereocystis and Macrocystis, providing the first direct hyperspectral comparison between the structures that make up the canopies of these species. Additionally, we compare the strength of their red-edge and near-infrared band signals, as well as the normalized difference red-edge (NDRE) and normalized difference vegetation index (NDVI) values. At the bed level, we compare detection of kelp canopy extent using both NDRE and NDVI classified unoccupied aerial vehicle imagery. We also characterized how changing tides and currents submerge the canopies of both species, providing insights that will allow remote sensors to more accurately determine the extent of kelp canopy in remote sensing imagery. Observations of canopy structures paired with in situ hyperspectral data and simulated multispectral data showed that more buoyant kelp structures had higher reflectance in the near-infrared wavelengths, but even slightly submerged canopy structures had a higher reflectance in the red-edge rather than the near-infrared. The higher red-edge signal was also evident at the bed level in the UAV imagery, resulting in 18.0% more canopy classified with NDRE than with NDVI. The area of detected canopy extent decreased by an average of 22.5% per meter of tidal increase at low current speeds (<10 cm/s), regardless of the species present. However, at higher current speeds (up to 19 cm/s), Nereocystis canopy decreased at nearly twice the average rate of kelp beds in low-current conditions. Apart from the strong differences in high-current regions, a robust linear relationship exists between kelp canopy extent and tidal height, which can aid in understanding the errors associated with remote sensing imagery collected at different tidal heights.

Assessing relationships of cover crop biomass and nitrogen content to multispectral imagery

AbstractCover crops provide valuable roles in sustainable agriculture, provided they produce enough biomass. To accurately measure their services to field management, spatial estimates would be useful to producers. This study used multispectral drone imagery to produce maps of normalized difference vegetation index (NDVI), normalized difference red edge index (NDRE), and a digital surface model (DSM) of cover crop plots on sandy, Mid‐Atlantic soils. Cover crops included cereal rye (Secale cereale), mixtures of rye and crimson clover (Trifolium incarnatum), and mixtures of rye and hairy vetch (Vicia villosa). Their biomass was sampled in the spring of 2019, 2020, and 2021, dried, weighed, and analyzed for total nitrogen (N) content. Measurements of NDVI became saturated (i.e., reached a linear plateau) at 3.86 Mg biomass ha−1, NDRE at 5.72 Mg biomass ha−1, and the DSM at 5.11 Mg biomass ha−1. The measured N content became saturated at 80.9, 139.1, and 75 kg N ha−1 for NDVI, NDRE, and the DSM, respectively. Based on log transformations, NDVI was a stronger predictor of biomass and N, but not C:N. The NDRE was important for biomass, N, and C:N, while the DSM interactions with cover crop species helped predict both the N content and C:N of cover crop tissues. Accumulated growing degree days was important as an individual variable for biomass and N and as an interaction with cover crop species.

UAV-Based Vegetation Indices to Evaluate Coffee Crop Response after Transplanting Seedlings Grown in Different Containers

Brazil stands out among coffee-growing countries worldwide. The use of precision agriculture to monitor coffee plants after transplantation has become an important step in the coffee production chain. The objective of this study was to assess how coffee plants respond after transplanting seedlings grown in different containers, based on multispectral images acquired by Unmanned Aerial Vehicles (UAV). The study was conducted in Santo Antônio do Amparo, Minas Gerais, Brazil. The coffee plants were imaged by UAV, and their height, crown diameter, and chlorophyll content were measured in the field. The vegetation indices were compared to the field measurements through graphical and correlation analysis. According to the results, no significant differences were found between the studied variables. However, the area transplanted with seedlings grown in perforated bags showed a lower percentage of mortality than the treatment with root trainers (6.4% vs. 11.7%). Additionally, the vegetation indices, including normalized difference red-edge, normalized difference vegetation index, and canopy planar area calculated by vectorization (cm2), were strongly correlated with biophysical parameters. Linear models were successfully developed to predict biophysical parameters, such as the leaf area index. Moreover, UAV proved to be an effective tool for monitoring coffee using this approach.

Assessing Forest Species Diversity in Ghana’s Tropical Forest Using PlanetScope Data

This study utilized a remotely sensed dataset with a high spatial resolution of 3 m to predict species diversity in the Bobiri Forest Reserve (BFR), a moist semi-deciduous tropical forest in Ghana. We conducted a field campaign of tree species measurements to achieve this objective for species diversity estimation. Thirty-five field plots of 50 m × 20 m were established, and the most dominant tree species within the forest were identified. Other measurements, such as diameter at breast height (DBH ≥ 5 cm), tree height, and each plot’s GPS coordinates, were recorded. The following species diversity indices were estimated from the field measurements: Shannon–Wiener (H′), Simpson diversity index (D2), species richness (S), and species evenness (J′). The PlanetScope surface reflectance data at 3 m spatial resolution was acquired and preprocessed for species diversity prediction. The spectral/pixel information of all bands, except the coastal band, was extracted for further processing. Vegetation indices (VIs) (NDVI—normalized difference vegetation index, EVI—enhanced vegetation index, SRI—simple ratio index, SAVI—soil adjusted vegetation index, and NDRE—normalized difference red edge index) were also calculated from the spectral bands and their pixel value extracted. A correlation analysis was then performed between the spectral bands and VIs with the species diversity index. The results showed that spectral bands 6 (red) and 2 (blue) significantly correlated with the two main species diversity indices (S and H′) due to their influence on vegetation properties, such as canopy biomass and leaf chlorophyll content. Furthermore, we conducted a stepwise regression analysis to investigate the most important spectral bands to consider when estimating species diversity from the PlanetScope satellite data. Like the correlation results, bands 6 (red) and 2 (blue) were the most important bands to be considered for predicting species diversity. The model equations from the stepwise regression were used to predict tree species diversity. Overall, the study’s findings emphasize the relevance of remotely sensed data in assessing the ecological condition of protected areas, a tool for decision-making in biodiversity conservation.

Delineation of Soil Management Zones and Validation through the Vigour of a Fodder Crop

In Mediterranean farming systems, the semi-arid conditions and agricultural ecosystems have made site-specific management an important approach. This method aims to understand and handle the variability of soil properties and crop management, particularly through the utilization of geospatial information and accessible technology. Over three years in a 30 ha experimental field located in the Alentejo region (Portugal), crop establishment was monitored using data from soil apparent electrical conductivity (ECa), remote sensing (Sentinel-2), and in situ soil sampling. The procedure began with Step 1, involving the acquisition of soil spatial information and spatial interpolation. Subsequently, in Step 2, management zones (MZs) for soil characteristics were delineated using a combination of ECa measurements and soil analysis, and Step 3 spanned over three years of gathering meteorological data and crop remote sensing measurements. In Step 4, site-specific crop MZs were delineated by vegetation indexes (VIs). This article aims to increase the importance of in situ and remote assessments to more accurately identify areas with different productive potential. Results showed three MZs based on the percentage of sand, ECa, altimetry, exchangeable calcium, and exchangeable calcium properties, validated by crop VIs (Normalized Difference Vegetation Index (NDVI), Normalized Difference Red-Edge Index (NDRE), and Normalized Difference Moisture Index (NDMI)) over time. Although there are many sensorial techniques available for site-specific management, this paper emphasizes a methodology for the farmer to identify different MZs combining remote and in situ evaluations, supporting new opportunities for a more rational use of natural resources. Based on soil parameters, three site-specific management areas could be selected. NDMI was the index that best explained the MZs created according to soil properties.

Enhancing estimation of cover crop biomass using field-based high-throughput phenotyping and machine learning models.

Incorporating cover crops into cropping systems offers numerous potential benefits, including the reduction of soil erosion, suppression of weeds, decreased nitrogen requirements for subsequent crops, and increased carbon sequestration. The aboveground biomass (AGB) of cover crops strongly influences their performance in delivering these benefits. Despite the significance of AGB, a comprehensive field-based high-throughput phenotyping study to quantify AGB of multiple cover crops in the U.S. Midwest has not been found. This study presents a two-year field experiment carried out in Eastern Nebraska, USA, to estimate AGB of five different cover crop species [canola (Brassica napus L.), rye (Secale cereale L.), triticale (Triticale × Triticosecale L.), vetch (Vicia sativa L.), and wheat (Triticum aestivum L.)] using high-throughput phenotyping and Machine Learning (ML) models. Destructive AGB sampling was performed three times during each spring season in 2022 and 2023. An array of morphological, spectral, thermal, and environmental features from the sensors were utilized as feature inputs of ML models. Moderately strong linear correlations between AGB and the selected features were observed. Four ML models, namely Random Forests Regression (RFR), Support Vector Regression (SVR), Partial Least Squares Regression (PLSR), and Artificial Neural Network (ANN), were investigated. Among the four models, PLSR achieved the highest Coefficient of Determination (R2) of 0.84 and the lowest Root Mean Squared Error (RMSE) of 892 kg/ha (Normalized RMSE (NRMSE) = 8.87%), indicating that PLSR could be the most appropriate method for estimating AGB of multiple cover crop species. Feature importance analysis ranked spectral features like Normalized Difference Red Edge (NDRE), Solar-induced Fluorescence (SIF), Spectral Reflectance at 485 nm (R485), and Normalized Difference Vegetation Index (NDVI) as top model features using PLSR. When utilizing fewer feature inputs, ANN exhibited better prediction performance compared to other models. Using morphological and spectral parameters as input features alone led to a R2 of 0.80 and 0.77 for AGB prediction using ANN, respectively. This study demonstrated the feasibility of high-throughput phenotyping and ML techniques for accurately estimating AGB of multiple cover crop species. Further enhancement of model performance could be achieved through additional destructive sampling conducted across multiple locations and years.

Using UAV multispectral photography to discriminate plant species in a seep wetland of the Fynbos Biome

Wetlands harbour a wide range of vital ecosystems. Hence, mapping wetlands is essential to conserving the ecosystems that depend on them. However, the physical nature of wetlands makes fieldwork difficult and potentially erroneous. This study used multispectral UAV aerial photography to map ten wetland plant species in the Fynbos Biome in the Steenbras Nature Reserve. We developed a methodology that used K-Nearest Neighbour (KNN), Support Vector Machine (SVM), and Random Forest (RF) machine learning algorithms to classify ten wetland plant species using the preselected bands and spectral indices. The study identified Normalized green red difference index (NGRDI), Red Green (RG) index, Green, Log Red Edge (LogRE), Normalized Difference Red-Edge (NDRE), Chlorophyll Index Red-Edge (CIRE), Green Ratio Vegetation Index (GRVI), Normalized Difference Water Index (NDWI), Green Normalized Difference Vegetation Index (GNDVI) and Red as pertinent bands and indices for classifying wetland plant species in the Proteaceae, Iridaceae, Restionaceae, Ericaceae, Asteraceae and Cyperaceae families. The classification had an overall accuracy of 87.4% and kappa accuracy of 0.85. Thus, the findings are pertinent to understanding the spectral characteristics of these endemic species. The study demonstrates the potential for UAV-based remote sensing of these endemic species.

Using Aerial Spectral Indices to Determine Fertility Rate and Timing in Winter Wheat

Tiller density is indicative of the overall health of winter wheat (Triticum aestivum L.) and is used to determine in-season nitrogen (N) application. If tiller density exceeds 538 tillers per m2 at GS 25, then an N application at that stage is not needed, only at GS 30. However, it is often difficult to obtain an accurate representation of tiller density across an entire field. Normalized difference vegetative index (NDVI) and normalized difference red edge (NDRE) have been significantly correlated with tiller density when collected from the ground. With the advent of unmanned aerial vehicles (UAVs) and their ability to collect NDVI and NDRE from the air, this study was established to examine the relationship between NDVI, NDRE, and tiller density, and to verify whether N could be applied based on these two indices. From 2018 to 2020, research trials were established across Virginia to develop a model describing the relationship between aerial NDVI, aerial NDRE, and tiller density counted on the ground, in small plots. In 2021, the model was used to apply N in small plots at two locations, where the obtained grain yield was the same whether N was applied based on tiller density, NDVI, or NDRE. From 2022 to 2023, the model was applied at six locations across the state on large scale growers’ fields to compare the amount of GS 25 N recommended by tiller density, NDVI, and NDRE. At three locations, NDVI and NDRE recommended the same N rates as the tiller density method, while at two locations, NDVI and NDRE recommended less N than tiller density. At one location, NDVI and NDRE recommended more N than tiller density. However, across all six locations, there was no difference in grain yield whether N was applied based on tiller density, NDVI, or NDRE. This study indicated that UAV-based NDVI and NDRE are excellent proxies for tiller density determination, and can be used to accurately and economically apply N at GS 25 in winter wheat production.

Using UAV-Based Temporal Spectral Indices to Dissect Changes in the Stay-Green Trait in Wheat.

Stay-green (SG) in wheat is a beneficial trait that increases yield and stress tolerance. However, conventional phenotyping techniques limited the understanding of its genetic basis. Spectral indices (SIs) as non-destructive tools to evaluate crop temporal senescence provide an alternative strategy. Here, we applied SIs to monitor the senescence dynamics of 565 diverse wheat accessions from anthesis to maturation stages over 2 field seasons. Four SIs (normalized difference vegetation index, green normalized difference vegetation index, normalized difference red edge index, and optimized soil-adjusted vegetation index) were normalized to develop relative stay-green scores (RSGS) as the SG indicators. An RSGS-based genome-wide association study identified 47 high-confidence quantitative trait loci (QTL) harboring 3,079 single-nucleotide polymorphisms associated with SG and 1,085 corresponding candidate genes. Among them, 15 QTL overlapped or were adjacent to known SG-related QTL/genes, while the remaining QTL were novel. Notably, a set of favorable haplotypes of SG-related candidate genes such as TraesCS2A03G1081100, TracesCS6B03G0356400, and TracesCS2B03G1299500 are increasing following the Green Revolution, further validating the feasibility of the pipeline. This study provided a valuable reference for further quantitative SG and genetic research in diverse wheat panels.

Time Series from Sentinel-2 for Organic Durum Wheat Yield Prediction Using Functional Data Analysis and Deep Learning

Yield prediction is a crucial activity in scheduling agronomic operations and in informing the management and financial decisions of a wide range of stakeholders of the organic durum wheat supply chain. This research aims to develop a yield forecasting system by combining vegetation index time-series data from Sentinel-2 L2A time-series data, field-measured yields, and deep learning techniques. Remotely sensed data over a season could be, in general, noisy and characterized by a variable density due to weather conditions. This problem was mitigated using Functional Principal Component Analysis (FPCA). We obtained a functional representation of acquired data, and starting from this, we tried to apply deep learning to predict the crop yield. We used a Convolutional Neural Network (CNN) approach, starting from images that embed temporal and spectral dimensions. This representation does not require one to a priori select a vegetation index that, typically, is task-dependent. The results have been also compared with classical approaches as Partial Least Squares (PLS) on the main reference vegetation indexes such as the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Red Edge index (NDRE), considering both in-season and end-season scenarios. The obtained results show that the image-based representation of multi-spectral time series could be an effective method to estimate the yield, also, in the middle stage of cropping with R2 values greater than 0.83. The developed model could be used to estimate yield the neighbor fields characterized by similar setups in terms of the crop, variety, soil, and, of course, management.

Combining multiple UAV-Based indicators for wheat yield estimation, a case study from Germany

ABSTRACT Unmanned aircraft vehicles (UAV) are widely used for yield estimations in agricultural production. Many significant improvements have been made towards the usage of hyperspectral and thermal sensors. The practical application of these new techniques meanwhile has been limited by the cost of data collection and the complexities of data processing. The objective of this paper is to evaluate the effectiveness of wheat yield estimations based on integrating vegetation indices (VI), solar radiation and crop height (CH), all of which are characterized by lower cost of data collection and processing. The VIs, solar radiation and CH were calculated based on UAV-based multispectral images obtained from two separate plots in Southern Germany and validated with data from a third plot. We compare the individual and joint predictive performance of different VIs, CH, and solar radiation by contrasting the estimated yield with actual yield based on multiple linear regression and quantile regression. The best predictive power was found for a combined estimation with CH, solar radiation and a Normalized Difference Red-edge Index (R2 = 0.75, RMSE = 0.53). This combined estimation resulted in a 15–20% improvement in the prediction of wheat yield accuracy as compared with utilizing any of the indices separately.

Pre-Harvest Corn Grain Moisture Estimation Using Aerial Multispectral Imagery and Machine Learning Techniques

Corn grain moisture (CGM) is critical to estimate grain maturity status and schedule harvest. Traditional methods for determining CGM range from manual scouting, destructive laboratory analyses, and weather-based dry down estimates. Such methods are either time consuming, expensive, spatially inaccurate, or subjective, therefore they are prone to errors or limitations. Realizing that precision harvest management could be critical for extracting the maximum crop value, this study evaluates the estimation of CGM at a pre-harvest stage using high-resolution (1.3 cm/pixel) multispectral imagery and machine learning techniques. Aerial imagery data were collected in the 2022 cropping season over 116 experimental corn planted plots. A total of 24 vegetation indices (VIs) were derived from imagery data along with reflectance (REF) information in the blue, green, red, red-edge, and near-infrared imaging spectrum that was initially evaluated for inter-correlations as well as subject to principal component analysis (PCA). VIs including the Green Normalized Difference Index (GNDVI), Green Chlorophyll Index (GCI), Infrared Percentage Vegetation Index (IPVI), Simple Ratio Index (SR), Normalized Difference Red-Edge Index (NDRE), and Visible Atmospherically Resistant Index (VARI) had the highest correlations with CGM (r: 0.68–0.80). Next, two state-of-the-art statistical and four machine learning (ML) models (Stepwise Linear Regression (SLR), Partial Least Squares Regression (PLSR), Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF), and K-nearest neighbor (KNN)), and their 120 derivates (six ML models × two input groups (REFs and REFs+VIs) × 10 train–test data split ratios (starting 50:50)) were formulated and evaluated for CGM estimation. The CGM estimation accuracy was impacted by the ML model and train-test data split ratio. However, the impact was not significant for the input groups. For validation over the train and entire dataset, RF performed the best at a 95:5 split ratio, and REFs+VIs as the input variables (rtrain: 0.97, rRMSEtrain: 1.17%, rentire: 0.95, rRMSEentire: 1.37%). However, when validated for the test dataset, an increase in the train–test split ratio decreased the performances of the other ML models where SVM performed the best at a 50:50 split ratio (r = 0.70, rRMSE = 2.58%) and with REFs+VIs as the input variables. The 95:5 train–test ratio showed the best performance across all the models, which may be a suitable ratio for relatively smaller or medium-sized datasets. RF was identified to be the most stable and consistent ML model (r: 0.95, rRMSE: 1.37%). Findings in the study indicate that the integration of aerial remote sensing and ML-based data-run techniques could be useful for reliably predicting CGM at the pre-harvest stage, and developing precision corn harvest scheduling and management strategies for the growers.

A comprehensive study on ecological insights of Ulva lactuca seaweed bloom in a lagoon along the southeast coast of India

Ulva lactuca, commonly known as sea lettuce is a green macroalgae (chlorophyta) and its uncontrolled proliferation forms a bloom, which causes ecological imbalances and adversely impacts brackishwater ecosystems. Algal bloom can harm the environment and aquatic organisms, but it can also be beneficial if harvested and utilized for industrial purposes. A massive bloom of green macroalgae was observed in the Muttukadu lagoon, located along the southeast coast of India in the Bay of Bengal. A comprehensive study was taken up to assess the spatial distribution, proximate, nutrients, minerals, heavy metals, and factors inducing the bloom. For this study, high-resolution satellite images (Sentinel 2) were used. The spatial distribution of seaweed was mapped using Floating Algae Index (FAI), Modified Floating Algae Index (MFAI), Normalized Difference Red Edge (NDRE) Index and Seaweed Index (SI) models. Results were validated using the MAVIC 3 pro drone and field survey. The SI model was found to be more accurate compared to other models. The bloom exhibited notable temporal dynamics, with peak coverage observed in the last week of May 2023 with a spread area of about 100.05 ha, the bloom intensity was found to be high till the last week of June 2023, after which it gradually declined. The proximate composition (%) of Ulva lactuca showed the moisture content is about 80.00 ± 2.98, crude protein 24.17 ± 1.80, crude lipid 0.81 ± 0.05, Total Ash 37.45 ± 1.44, Fiber (8.92 ± 0.41), and Acid Insoluble Ash (16.18 ± 1.16). The mineral concentration was in the order Ca > Na > Mg > K > P. Iron (Fe) was observed to be high compared to other micronutrients, with a concentration of 1164.40 mg/kg. Rich nutrients and low heavy metal concentration significantly can boost the selling cost of Ulva lactuca providing livelihood support for the coastal fisherfolk.

IMPACT OF UAV AND SENTINEL-2A IMAGERY FUSION ON VEGETATION INDICES PERFORMANCE

Abstract. Image fusion techniques can improve the quality of remote sensing images by combining high spatial resolution images with low spectral resolution images. This enhancement of the images may impact the performance of various vegetation indices (VI’s). This study investigates the impact of image fusion on the quality of vegetation indices by fusing UAV (Unmanned Aerial Vehicle) bands with Sentinel-2A images using Principal Component Analysis (PCA) and Brovery Transform (BT) fusion techniques.The fused images were used to calculate the Normalized Difference Vegetation Index, Normalized Difference Red Edge, Green Red Vegetation Index, and Normalized Difference Water Index. To assess the performance of the fused images, several image quality assessment metrics were used, including Root Mean Square Error (RMSE), Entropy, etc...The results showed that image fusion techniques can improve the quality of images which is important to assess crop health. The PCA image fusion technique showed higher quality than the BT technique. The PCA fused images had lower RMSE, ERGAS, and Entropy Difference and higher UIQI, CC, and SSIM values than the original images. Moreover, the fused images produced higher VIs values than the Sentinel-2A images.Finally, scatter plots were created to compare the correlation between the VIs calculated from the original and fused images. The results showed a strong correlation between the VIs calculated from the Sentinel-2A and fused images, indicating that the fused images can accurately estimate vegetation health parameters. Overall, this study demonstrates the potential of image fusion techniques to improve the quality of VI’s for monitoring vegetation health.