County-Level Soybean Yield Prediction Using Deep CNN-LSTM Model

Integrating multi-modal remote sensing, deep learning, and attention mechanisms for yield prediction in plant breeding experiments

In both plant breeding and crop management, interpretability plays a crucial role in instilling trust in AI-driven approaches and enabling the provision of actionable insights. The primary objective of this research is to explore and evaluate the potential contributions of deep learning network architectures that employ stacked LSTM for end-of-season maize grain yield prediction. A secondary aim is to expand the capabilities of these networks by adapting them to better accommodate and leverage the multi-modality properties of remote sensing data. In this study, a multi-modal deep learning architecture that assimilates inputs from heterogeneous data streams, including high-resolution hyperspectral imagery, LiDAR point clouds, and environmental data, is proposed to forecast maize crop yields. The architecture includes attention mechanisms that assign varying levels of importance to different modalities and temporal features that, reflect the dynamics of plant growth and environmental interactions. The interpretability of the attention weights is investigated in multi-modal networks that seek to both improve predictions and attribute crop yield outcomes to genetic and environmental variables. This approach also contributes to increased interpretability of the model's predictions. The temporal attention weight distributions highlighted relevant factors and critical growth stages that contribute to the predictions. The results of this study affirm that the attention weights are consistent with recognized biological growth stages, thereby substantiating the network's capability to learn biologically interpretable features. Accuracies of the model's predictions of yield ranged from 0.82-0.93 R2 ref in this genetics-focused study, further highlighting the potential of attention-based models. Further, this research facilitates understanding of how multi-modality remote sensing aligns with the physiological stages of maize. The proposed architecture shows promise in improving predictions and offering interpretable insights into the factors affecting maize crop yields, while demonstrating the impact of data collection by different modalities through the growing season. By identifying relevant factors and critical growth stages, the model's attention weights provide valuable information that can be used in both plant breeding and crop management. The consistency of attention weights with biological growth stages reinforces the potential of deep learning networks in agricultural applications, particularly in leveraging remote sensing data for yield prediction. To the best of our knowledge, this is the first study that investigates the use of hyperspectral and LiDAR UAV time series data for explaining/interpreting plant growth stages within deep learning networks and forecasting plot-level maize grain yield using late fusion modalities with attention mechanisms.

Use of one-dimensional CNN for input data size reduction in LSTM for improved computational efficiency and accuracy in hourly rainfall-runoff modeling

A deep learning architecture, denoted as CNNsLSTM, is proposed for hourly rainfall-runoff modeling in this study. The architecture involves a serial coupling of the one-dimensional convolutional neural network (1D-CNN) and the long short-term memory (LSTM) network. In the proposed framework, multiple layers of the CNN component process long-term hourly meteorological time series data, while the LSTM component handles short-term meteorological time series data and utilizes the extracted features from the 1D-CNN. In order to demonstrate the effectiveness of the proposed approach, it was implemented for hourly rainfall-runoff modeling in the Ishikari River watershed, Japan. A meteorological dataset, including precipitation, air temperature, evapotranspiration, longwave radiation, and shortwave radiation, was utilized as input. The results of the proposed approach (CNNsLSTM) were compared with those of previously proposed deep learning approaches used in hydrologic modeling, such as 1D-CNN, LSTM with only hourly inputs (LSTMwHour), a parallel architecture of 1D-CNN and LSTM (CNNpLSTM), and the LSTM architecture, which uses both daily and hourly input data (LSTMwDpH). Meteorological and runoff datasets were separated into training, validation, and test periods to train the deep learning model without overfitting, and evaluate the model with an independent dataset. The proposed approach clearly improved estimation accuracy compared to previously utilized deep learning approaches in rainfall = runoff modeling. In comparison with the observed flows, the median values of the Nash-Sutcliffe efficiency for the test period were 0.455-0.469 for 1D-CNN, 0.639-0.656 for CNNpLSTM, 0.745 for LSTMwHour, 0.831 for LSTMwDpH, and 0.865-0.873 for the proposed CNNsLSTM. Furthermore, the proposed CNNsLSTM reduced the median root mean square error (RMSE) of 1D-CNN by 50.2%-51.4%, CNNpLSTM by 37.4%-40.8%, LSTMwHour by 27.3%-29.5%, and LSTMwDpH by 10.6%-13.4%. Particularly, the proposed CNNsLSTM improved the estimations for high flows (≧75th percentile) and peak flows (≧95th percentile). The computational speed of LSTMwDpH is the fastest among the five architectures. Although the computation speed of CNNsLSTM is slower than LSTMwDpH's, it is still 6.9-7.9 times faster than that of LSTMwHour. Therefore, the proposed CNNsLSTM would be an effective approach for flood management and hydraulic structure design, mainly under climate change conditions that require estimating hourly river flows using meteorological datasets.

In-Season Cotton Yield Prediction with Scale-Aware Convolutional Neural Network Models and Unmanned Aerial Vehicle RGB Imagery

In the pursuit of sustainable agriculture, efficient water management remains crucial, with growers relying on advanced techniques for informed decision-making. Cotton yield prediction, a critical aspect of agricultural planning, benefits from cutting-edge technologies. However, traditional methods often struggle to capture the nuanced complexities of crop health and growth. This study introduces a novel approach to cotton yield prediction, leveraging the synergy between Unmanned Aerial Vehicles (UAVs) and scale-aware convolutional neural networks (CNNs). The proposed model seeks to harness the spatiotemporal dynamics inherent in high-resolution UAV imagery to improve the accuracy of the cotton yield prediction. The CNN component adeptly extracts spatial features from UAV-derived imagery, capturing intricate details related to crop health and growth, modeling temporal dependencies, and facilitating the recognition of trends and patterns over time. Research experiments were carried out in a cotton field at the USDA-ARS Cropping Systems Research Laboratory (CSRL) in Lubbock, Texas, with three replications evaluating four irrigation treatments (rainfed, full irrigation, percent deficit of full irrigation, and time delay of full irrigation) on cotton yield. The prediction revealed that the proposed CNN regression models outperformed conventional CNN models, such as AlexNet, CNN-3D, CNN-LSTM, ResNet. The proposed CNN model showed state-of-art performance at different image scales, with the R2 exceeding 0.9. At the cotton row level, the mean absolute error (MAE) and mean absolute percentage error (MAPE) were 3.08 pounds per row and 7.76%, respectively. At the cotton grid level, the MAE and MAPE were 0.05 pounds and 10%, respectively. This shows the proposed model's adaptability to the dynamic interplay between spatial and temporal factors that affect cotton yield. The authors conclude that integrating UAV-derived imagery and CNN regression models is a potent strategy for advancing precision agriculture, providing growers with a powerful tool to optimize cultivation practices and enhance overall cotton productivity.

GOA-optimized deep learning for soybean yield estimation using multi-source remote sensing data

Accurately estimating large-area crop yields, especially for soybeans, is essential for addressing global food security challenges. This study introduces a deep learning framework that focuses on precise county-level soybean yield estimation in the United States. It utilizes a wide range of multi-variable remote sensing data. The model used in this study is a state-of-the-art CNN-BiGRU model, which is enhanced by the GOA and a novel attention mechanism (GCBA). This model excels in handling intricate time series and diverse remote sensing datasets. Compared to five leading machine learning and deep learning models, our GCBA model demonstrates superior performance, particularly in the 2019 and 2020 evaluations, achieving remarkable R2, RMSE, MAE and MAPE values. This sets a new benchmark in yield estimation accuracy. Importantly, the study highlights the significance of integrating multi-source remote sensing data. It reveals that synthesizing information from various sensors and incorporating photosynthesis-related parameters significantly enhances yield estimation precision. These advancements not only provide transformative insights for precision agricultural management but also establish a solid scientific foundation for informed decision-making in global agricultural production and food security.

Combining machine learning and remote sensing-integrated crop modeling for rice and soybean crop simulation

Machine learning (ML) techniques offer a promising avenue for improving the integration of remote sensing data into mathematical crop models, thereby enhancing crop growth prediction accuracy. A critical variable for this integration is the leaf area index (LAI), which can be accurately assessed using proximal or remote sensing data based on plant canopies. This study aimed to (1) develop a machine learning-based method for estimating the LAI in rice and soybean crops using proximal sensing data and (2) evaluate the performance of a Remote Sensing-Integrated Crop Model (RSCM) when integrated with the ML algorithms. To achieve these objectives, we analyzed rice and soybean datasets to identify the most effective ML algorithms for modeling the relationship between LAI and vegetation indices derived from canopy reflectance measurements. Our analyses employed a variety of ML regression models, including ridge, lasso, support vector machine, random forest, and extra trees. Among these, the extra trees regression model demonstrated the best performance, achieving test scores of 0.86 and 0.89 for rice and soybean crops, respectively. This model closely replicated observed LAI values under different nitrogen treatments, achieving Nash-Sutcliffe efficiencies of 0.93 for rice and 0.97 for soybean. Our findings show that incorporating ML techniques into RSCM effectively captures seasonal LAI variations across diverse field management practices, offering significant potential for improving crop growth and productivity monitoring.

Impacts of meteorological variables and machine learning algorithms on rice yield prediction in Korea

As crop productivity is greatly influenced by weather conditions, many attempts have been made to estimate crop yields using meteorological data and have achieved great progress with the development of machine learning. However, most yield prediction models are developed based on observational data, and the utilization of climate model output in yield prediction has been addressed in very few studies. In this study, we estimate rice yields in South Korea using the meteorological variables provided by ERA5 reanalysis data (ERA-O) and its dynamically downscaled data (ERA-DS). After ERA-O and ERA-DS are validated against observations (OBS), two different machine learning models, Support Vector Machine (SVM) and Long Short-Term Memory (LSTM), are trained with different combinations of eight meteorological variables (mean temperature, maximum temperature, minimum temperature, precipitation, diurnal temperature range, solar irradiance, mean wind speed, and relative humidity) obtained from OBS, ERA-O, and ERA-DS at weekly and monthly timescales from May to September. Regardless of the model type and the source of the input data, training a model with weekly datasets leads to better yield estimates compared to monthly datasets. LSTM generally outperforms SVM, especially when the model is trained with ERA-DS data at a weekly timescale. The best yield estimates are produced by the LSTM model trained with all eight variables at a weekly timescale. Altogether this study shows the significance of high spatial and temporal resolution of input meteorological data in yield prediction, which can also serve to substantiate the added value of dynamical downscaling.

Spatial and temporal characteristics analysis and prediction model of PM2.5 concentration based on SpatioTemporal-Informer model

The primary cause of hazy weather is PM2.5, and forecasting PM2.5 concentrations can aid in managing and preventing hazy weather. This paper proposes a novel spatiotemporal prediction model called SpatioTemporal-Informer (ST-Informer) in response to the shortcomings of spatiotemporal prediction models commonly used in studies for long-input series prediction. The ST-Informer model implements parallel computation of long correlations and adds an independent spatiotemporal embedding layer to the original Informer model. The spatiotemporal embedding layer captures the complex dynamic spatiotemporal correlations among the input information. In addition, the ProbSpare Self-Attention mechanism in this model can focus on extracting important contextual information of spatiotemporal data. The ST-Informer model uses weather and air pollutant concentration data from numerous stations as its input data. The outcomes of the trials indicate that (1) The ST-Informer model can sharply capture the peaks and sudden changes in PM2.5 concentrations. (2) Compared to the current models, the ST-Informer model shows better prediction performance while maintaining high-efficiency prediction [Formula: see text]. (3) The ST-Informer model has universal applicability, and the model was applied to the concentration of other pollutants prediction with good results.

A transformer-based approach for early prediction of soybean yield using time-series images

Crop yield prediction which provides critical information for management decision-making is of significant importance in precision agriculture. Traditional manual inspection and calculation are often laborious and time-consuming. For yield prediction using high-resolution images, existing methods, e.g., convolutional neural network, are challenging to model long range multi-level dependencies across image regions. This paper proposes a transformer-based approach for yield prediction using early-stage images and seed information. First, each original image is segmented into plant and soil categories. Two vision transformer (ViT) modules are designed to extract features from each category. Then a transformer module is established to deal with the time-series features. Finally, the image features and seed features are combined to estimate the yield. A case study has been conducted using a dataset that was collected during the 2020 soybean-growing seasons in Canadian fields. Compared with other baseline models, the proposed method can reduce the prediction error by more than 40%. The impact of seed information on predictions is studied both between models and within a single model. The results show that the influence of seed information varies among different plots but it is particularly important for the prediction of low yields.

Comparison of PlanetScope, Sentinel-2, and landsat 8 data in soybean yield estimation within-field variability with random forest regression

Accurate timely and early-season crop yield estimation within the field variability is important for precision farming and sustainable management applications. Therefore, the ability to estimate the within-field variability of grain yield is crucial for ensuring food security worldwide, especially under climate change. Several Earth observation systems have thus been developed to monitor crops and predict yields. Despite this, new research is required to combine multiplatform data integration, advancements in satellite technologies, data processing, and the application of this discipline to agricultural practices. This study provides further developments in soybean yield estimation by comparing multisource satellite data from PlanetScope (PS), Sentinel-2 (S2), and Landsat 8 (L8) and introducing topographic and meteorological variables. Herein, a new method of combining soybean yield, global positioning systems, harvester data, climate, topographic variables, and remote sensing images has been demonstrated. Soybean yield shape points were obtained from a combine-harvester-installed GPS and yield monitoring system from seven fields over the 2021 season. The yield estimation models were trained and validated using random forest, and four vegetation indices were tested. The result showed that soybean yield can be accurately predicted at 3-, 10-, and 30-m resolutions with mean absolute error (MAE) value of 0.091 t/ha for PS, 0.118 t/ha for S2, and 0.120 t/ha for L8 data (root mean square error (RMSE) of 0.111, 0.076). The combination of the environmental data with the original bands provided further improvements and an accurate yield estimation model within the soybean yield variability with MAE of 0.082 t/ha for PS, 0.097 t/ha for S2, and 0.109 t/ha for L8 (RMSE of 0.094, 0.069, and 0.108 t/ha). The results showed that the optimal date to predict the soybean yield within the field scale was approximately 60 or 70 days before harvesting periods during the beginning bloom stage. The developed model can be applied for other crops and locations when suitable training yield data, which are critical for precision farming, are available.

A data-driven crop model for maize yield prediction

Accurate estimation of crop yield predictions is of great importance for food security under the impact of climate change. We propose a data-driven crop model that combines the knowledge advantage of process-based modeling and the computational advantage of data-driven modeling. The proposed model tracks the daily biomass accumulation process during the maize growing season and uses daily produced biomass to estimate the final grain yield. Computational studies using crop yield, field location, genotype and corresponding environmental data were conducted in the US Corn Belt region from 1981 to 2020. The results suggest that the proposed model can achieve an accurate prediction performance with a 7.16% relative root-mean-square-error of average yield in 2020 and provide scientifically explainable results. The model also demonstrates its ability to detect and separate interactions between genotypic parameters and environmental variables. Additionally, this study demonstrates the potential value of the proposed model in helping farmers achieve higher yields by optimizing seed selection.

Environmental Prediction in Cold Chain Transportation of Agricultural Products Based on K-Means++ and LSTM Neural Network

Experiments have proven that traditional prediction research methods have limitations in practice. Proposing countermeasures for environmental changes is the key to optimal control of the cold chain environment and reducing the lag of control effects. In this paper, a cold chain transportation environment prediction method, combining k-means++ and a long short-term memory (LSTM) neural network, is proposed according to the characteristics of the cold chain transportation environment of agricultural products. The proposed prediction model can predict the trend of cold chain environment changes in the next ten minutes, which allows cold chain vehicle managers to issue control instructions to the environmental control equipment in advance. The fusion process for temperature and humidity data measured by multiple data sensors is performed with the k-means++ algorithm, and then the fused data are fed into an LSTM neural network for prediction based on time series. The prediction error of the prediction model proposed in this paper is very satisfactory, with a root-mean-square error (RMSE), mean absolute error (MAE), mean squared error (MSE), mean absolute percentage error (MAPE) and R-squared of 0.5707, 0.2484, 0.3258, 0.0312 and 0.9660, respectively, for temperature prediction, and with an RMSE, MAE, MSE, mean absolute percentage error and R-squared of 1.6015, 1.1770, 2.5648, 0.2736 and 0.9702, respectively, for humidity prediction. Finally, the LSTM neural network and back propagation (BP) neural network are compared in order to enhance the reliability of the results. In terms of the prediction effect of the temperature and humidity in cold chain vehicles transporting agricultural products, the proposed model has a higher prediction accuracy than that of existing models and can provide strategic support for the fine management and regulation of the cold chain transportation environment.

Integrating environmental and satellite data to estimate county-level cotton yield in Xinjiang Province

Accurate and timely estimation of cotton yield over large areas is essential for precision agriculture, facilitating the operation of commodity markets and guiding agronomic management practices. Remote sensing (RS) and crop models are effective means to predict cotton yield in the field. The satellite vegetation indices (VIs) can describe crop yield variations over large areas but can't take the exact environmental impact into consideration. Climate variables (CVs), the result of the influence of spatial heterogeneity in large regions, can provide environmental information for better estimation of cotton yield. In this study, the most important VIs and CVs for estimating county-level cotton yield across Xinjiang Province were screened out. We found that the VIs of canopy structure and chlorophyll contents, and the CVs of moisture, were the most significant factors for cotton growth. For yield estimation, we utilized four approaches: least absolute shrinkage and selection operator regression (LASSO), support vector regression (SVR), random forest regression (RFR) and long short-term memory (LSTM). Due to its ability to capture temporal features over the long term, LSTM performed best, with an R2 of 0.76, root mean square error (RMSE) of 150 kg/ha and relative RMSE (rRMSE) of 8.67%; moreover, an additional 10% of the variance could be explained by adding CVs to the VIs. For the within-season yield estimation using LSTM, predictions made 2 months before harvest were the most accurate (R2 = 0.65, RMSE = 220 kg/ha, rRMSE = 15.97%). Our study demonstrated the feasibility of yield estimation and early prediction at the county level over large cotton cultivation areas by integrating satellite and environmental data.

A Data-Driven OBE Magnetic Interference Compensation Method

Aeromagnetic compensation is a technology used to reduce aircraft magnetic interference, which plays an important role in aeromagnetic surveys. In addition to maneuvering interferences, the onboard electronic (OBE) interference has been proven to be a significant part of aircraft interference, which must be reduced before further interpretation of aeromagnetic data. In the past, most researchers have focused on establishing linear models to compensate for OBE magnetic interference. However, such methods can only work using accurate reference sensors. In this paper, we propose a data-driven OBE interference compensation method, which can reduce OBE interference without relying on any other reference sensor. This network-based method can integrally detect and repair the OBE magnetic interference. The proposed method builds a prediction model by combining wavelet decomposition with a long short-term memory (LSTM) network to detect and predict OBE interference, and then estimates the local variation of the magnetic field to remove the drift of the interference. In our tests, we construct 10 semi-real datasets to quantitatively evaluate the performance of the proposed method. The F1 score of the proposed method for OBE interference detection is over 0.79, and the RMSE of the compensated signal is less than 0.009 nT. Moreover, we also test our method on real signals, and the results show that our method can detect all interference and significantly reduce the standard deviation of the magnetic field.

Soybean cyst nematode detection and management: a review

Soybeans play a key role in global food security. U.S. soybean yields, which comprise [Formula: see text] of the total soybeans planted in the world, continue to experience unprecedented grain loss due to the soybean cyst nematode (SCN) plant pathogen. SCN remains one of the primary disruptive pests despite the existence of advanced management techniques such as crop rotation and SCN-resistant varieties. SCN detection is a key step in managing this disease; however, early detection is challenging because soybeans do not show any above ground symptoms unless they  are significantly damaged. Direct soil sampling remains the most common method for SCN detection, however, this method has several problems. For example, the threshold damage methods-adopted by most of the laboratories to make recommendations-is not reliable as it does not consider soil pH, N, P, and K values and relies solely on the egg count instead of assessment of the root infection. To overcome the challenges of manual soil sampling methods, deep learning and hyperspectral imaging are important current topics in precision agriculture for plant disease detection and have been proposed as cost-effective and efficient detection methods that can work at scale. We have reviewed more than 150 research papers focusing on soybean cyst nematodes with an emphasis on deep learning techniques for detection and management. First: we describe soybean vegetation and reproduction stages, SCN life cycles, and factors influencing this disease. Second: we highlight the impact of SCN on soybean yield loss and the challenges associated with its detection. Third: we describe direct sampling methods in which the soil samples are procured and analyzed to evaluate SCN egg counts. Fourth: we highlight the advantages and limitations of these direct methods, then review computer vision- and remote sensing-based detection methods: data collection using ground, aerial, and satellite approaches followed by a review of machine learning methods for image analysis-based soybean cyst nematode detection. We highlight the evaluation approaches and the advantages of overall detection workflow in high-performance and big data environments. Lastly, we discuss various management approaches, such as crop rotation, fertilization, SCN resistant varieties such as PI 88788, and SCN's increasing resistance to these strategies. We review machine learning approaches for soybean crop yield forecasting as well as the influence of pesticides, herbicides, and fertilizers on SCN infestation reduction. We provide recommendations for soybean research using deep learning and hyperspectral imaging to accommodate the lack of the ground truth data and training and testing methodologies, such as data augmentation and transfer learning, to achieve a high level of detection accuracy while keeping costs as low as possible.

An Intelligent Rice Yield Trait Evaluation System Based on Threshed Panicle Compensation

High-throughput phenotyping of yield-related traits is meaningful and necessary for rice breeding and genetic study. The conventional method for rice yield-related trait evaluation faces the problems of rice threshing difficulties, measurement process complexity, and low efficiency. To solve these problems, a novel intelligent system, which includes an integrated threshing unit, grain conveyor-imaging units, threshed panicle conveyor-imaging unit, and specialized image analysis software has been proposed to achieve rice yield trait evaluation with high throughput and high accuracy. To improve the threshed panicle detection accuracy, the Region of Interest Align, Convolution Batch normalization activation with Leaky Relu module, Squeeze-and-Excitation unit, and optimal anchor size have been adopted to optimize the Faster-RCNN architecture, termed 'TPanicle-RCNN,' and the new model achieved F1 score 0.929 with an increase of 0.044, which was robust to indica and japonica varieties. Additionally, AI cloud computing was adopted, which dramatically reduced the system cost and improved flexibility. To evaluate the system accuracy and efficiency, 504 panicle samples were tested, and the total spikelet measurement error decreased from 11.44 to 2.99% with threshed panicle compensation. The average measuring efficiency was approximately 40 s per sample, which was approximately twenty times more efficient than manual measurement. In this study, an automatic and intelligent system for rice yield-related trait evaluation was developed, which would provide an efficient and reliable tool for rice breeding and genetic research.

Google Earth Engine and Artificial Intelligence (AI): A Comprehensive Review

Remote sensing (RS) plays an important role gathering data in many critical domains (e.g., global climate change, risk assessment and vulnerability reduction of natural hazards, resilience of ecosystems, and urban planning). Retrieving, managing, and analyzing large amounts of RS imagery poses substantial challenges. Google Earth Engine (GEE) provides a scalable, cloud-based, geospatial retrieval and processing platform. GEE also provides access to the vast majority of freely available, public, multi-temporal RS data and offers free cloud-based computational power for geospatial data analysis. Artificial intelligence (AI) methods are a critical enabling technology to automating the interpretation of RS imagery, particularly on object-based domains, so the integration of AI methods into GEE represents a promising path towards operationalizing automated RS-based monitoring programs. In this article, we provide a systematic review of relevant literature to identify recent research that incorporates AI methods in GEE. We then discuss some of the major challenges of integrating GEE and AI and identify several priorities for future research. We developed an interactive web application designed to allow readers to intuitively and dynamically review the publications included in this literature review.

Machine learning algorithms for soybean yield forecasting in the Brazilian Cerrado

BACKGROUND

We evaluated different machine learning (ML) models for predicting soybean productivity up to 1 month in advance for the Matopiba agricultural frontier (States of Maranhão, Tocantins, Piauí, and Bahia). We collected meteorological data on the NASA-POWER platform and soybean yield on the SIDRA/IBGE base between 2008 and 2017. The ML models evaluated were random forest (RF), artificial neural networks, radial base support vector machines (SVM_RBF), linear model and polynomial regression. To assess the performance of the models, cross-validation was used, obtaining the value of precision by R2 , accuracy by root mean square error (RMSE), and trend by the mean error of the estimate (EME).

RESULTS

The results showed that the RF algorithm achieves the highest precision and accuracy, with R2 of 0.81, RMSE of 176.93 kg ha-1 and trend (EME) of 1.99 kg ha-1 . On the other hand, the SVM_RBF algorithm showed the lowest performance, with R2 of 0.74, RMSE of 213.58 kg ha-1 and EME of -15.06 kg ha-1 . The average yield values predicted by the models were within the expected range for the region, which has a historical average value of 2.730 kg ha-1 .

CONCLUSION

All models had acceptable precision, accuracy and trend indices, which makes it possible to use all algorithms to be applied in the prediction of soybean crop yield, observing the particularities of the region to be studied, in addition to being a useful tool for agricultural planning and decision making in soy-producing regions such as the Brazilian Cerrado. © 2021 Society of Chemical Industry.

LSTM-Based Prediction of Mediterranean Vegetation Dynamics Using NDVI Time-Series Data

Vegetation index time-series analysis of multitemporal satellite data is widely used to study vegetation dynamics in the present climate change era. This paper proposes a systematic methodology to predict the Normalized Difference Vegetation Index (NDVI) using time-series data extracted from the Moderate Resolution Imaging Spectroradiometer (MODIS). The key idea is to obtain accurate NDVI predictions by combining the merits of two effective computational intelligence techniques; namely, fuzzy clustering and long short-term memory (LSTM) neural networks under the framework of dynamic time warping (DTW) similarity measure. The study area is the Lesvos Island, located in the Aegean Sea, Greece, which is an insular environment in the Mediterranean coastal region. The algorithmic steps and the main contributions of the current work are described as follows. (1) A data reduction mechanism was applied to obtain a set of representative time series. (2) Since DTW is a similarity measure and not a distance, a multidimensional scaling approach was applied to transform the representative time series into points in a low-dimensional space, thus enabling the use of the Euclidean distance. (3) An efficient optimal fuzzy clustering scheme was implemented to obtain the optimal number of clusters that better described the underline distribution of the low-dimensional points. (4) The center of each cluster was mapped into time series, which were the mean of all representative time series that corresponded to the points belonging to that cluster. (5) Finally, the time series obtained in the last step were further processed in terms of LSTM neural networks. In particular, development and evaluation of the LSTM models was carried out considering a one-year period, i.e., 12 monthly time steps. The results indicate that the method identified unique time-series patterns of NDVI among different CORINE land-use/land-cover (LULC) types. The LSTM networks predicted the NDVI with root mean squared error (RMSE) ranging from 0.017 to 0.079. For the validation year of 2020, the difference between forecasted and actual NDVI was less than 0.1 in most of the study area. This study indicates that the synergy of the optimal fuzzy clustering based on DTW similarity of NDVI time-series data and the use of LSTM networks with clustered data can provide useful results for monitoring vegetation dynamics in fragmented Mediterranean ecosystems.

A Geographically Weighted Random Forest Approach to Predict Corn Yield in the US Corn Belt

Crop yield prediction before the harvest is crucial for food security, grain trade, and policy making. Previously, several machine learning methods have been applied to predict crop yield using different types of variables. In this study, we propose using the Geographically Weighted Random Forest Regression (GWRFR) approach to improve crop yield prediction at the county level in the US Corn Belt. We trained the GWRFR and five other popular machine learning algorithms (Multiple Linear Regression (MLR), Partial Least Square Regression (PLSR), Support Vector Regression (SVR), Decision Tree Regression (DTR), and Random Forest Regression (RFR)) with the following different sets of features: (1) full length features; (2) vegetation indices; (3) gross primary production (GPP); (4) climate data; and (5) soil data. We compared the results of the GWRFR with those of the other five models. The results show that the GWRFR with full length features (R2 = 0.90 and RMSE = 0.764 MT/ha) outperforms other machine learning algorithms. For individual categories of features such as GPP, vegetation indices, climate, and soil features, the GWRFR also outperforms other models. The Moran’s I value of the residuals generated by GWRFR is smaller than that of other models, which shows that GWRFR can better address the spatial non-stationarity issue. The proposed method in this article can also be potentially used to improve yield prediction for other types of crops in other regions.

Research on Rice Yield Prediction Model Based on Deep Learning

Food is the paramount necessity of the people. With the progress of society and the improvement of social welfare system, the living standards of people all over the world are constantly improving. The development of medical industry improves people's health level constantly, and the world population is constantly climbing to a new peak. With the continuous development of deep learning in recent years, its advantages are constantly displayed, especially in the aspect of image recognition and processing, it drives into the distance. Thanks to the superiority of deep learning in image processing, the combination of remote sensing images and deep learning has attracted more attention. To simulate the four key factors of rice yield, this article tries a regression model with a combination of various characteristic independent variables. In this article, the selection of the best linear and nonlinear regression models is discussed, the prediction performance and significance of each regression model are analyzed, and some thoughts are given on estimation of actual rice yield.

Soya Yield Prediction on a Within-Field Scale Using Machine Learning Models Trained on Sentinel-2 and Soil Data

Agriculture is the backbone and the main sector of the industry for many countries in the world. Assessing crop yields is key to optimising on-field decisions and defining sustainable agricultural strategies. Remote sensing applications have greatly enhanced our ability to monitor and manage farming operation. The main objective of this research was to evaluate machine learning system for within-field soyayield prediction trained on Sentinel-2 multispectral images and soil parameters. Multispectral images used in the study came from ESA’s Sentinel-2 satellites. A total of 3 cloud-free Sentinel-2 multispectral images per year from specific periods of vegetation were used to obtain the time-series necessary for crop yield prediction. Yield monitor data were collected in three crop seasons (2018, 2019 and 2020) from a number of farms located in Upper Austria. The ground-truth database consisted of information about the location of the fields and crop yield monitor data on 411 ha of farmland. A novel method, namely the Polygon-Pixel Interpolation, for optimal fitting yield monitor data with satellite images is introduced. Several machine learning algorithms, such as Multiple Linear Regression, Support Vector Machine, eXtreme Gradient Boosting, Stochastic Gradient Descent and Random Forest, were compared for their performance in soya yield prediction. Among the tested machine learning algorithms, Stochastic Gradient Descent regression model performed better than the others, with a mean absolute error of 4.36 kg/pixel (0.436 t/ha) and a correlation coefficient of 0.83%.

A Systematic Literature Review on Crop Yield Prediction with Deep Learning and Remote Sensing

Deep learning has emerged as a potential tool for crop yield prediction, allowing the model to automatically extract features and learn from the datasets. Meanwhile, smart farming technology enables the farmers to achieve maximum crop yield by extracting essential parameters of crop growth. This systematic literature review highlights the existing research gaps in a particular area of deep learning methodologies and guides us in analyzing the impact of vegetation indices and environmental factors on crop yield. To achieve the aims of this study, prior studies from 2012 to 2022 from various databases are collected and analyzed. The study focuses on the advantages of using deep learning in crop yield prediction, the suitable remote sensing technology based on the data acquisition requirements, and the various features that influence crop yield prediction. This study finds that Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN) are the most widely used deep learning approaches for crop yield prediction. The commonly used remote sensing technology is satellite remote sensing technology—in particular, the use of the Moderate-Resolution Imaging Spectroradiometer (MODIS). Findings show that vegetation indices are the most used feature for crop yield prediction. However, it is also observed that the most used features in the literature do not always work for all the approaches. The main challenges of using deep learning approaches and remote sensing for crop yield prediction are how to improve the working model for better accuracy, the practical implication of the model for providing accurate information about crop yield to agriculturalists, growers, and policymakers, and the issue with the black box property.

A Novel Prediction and Planning Model for the Benefit of Irrigation Water Allocation Based on Deep Learning and Uncertain Programming

Due to population growth and human activities, water shortages have become an increasingly serious concern in the North China Plain, which has become the world’s largest underground water funnel. Because the yield per unit area, planting area of crops, and effective precipitation in the region are uncertain, it is not easy to plan the amount of irrigation water for crops. In order to improve the applicability of the uncertainty programming model, a hybrid LSTM-CPP-FPP-IPP model (long short-term memory, chance-constrained programming, fuzzy possibility programming, interval parameter programming) was developed to plan the irrigation water allocation of irrigation system under uncertainty. The LSTM (long short-term memory) model was used to predict crop yield per unit area, and CPP-FPP-IPP programming (chance-constrained programming, fuzzy possibility programming, interval parameter programming) was used to plan the crop area and the effective precipitation under uncertainty. The hybrid model was used for the crop production profit of winter wheat and summer corn in five cities in the North China Plain. The average absolute error between the model prediction value and the actual value of the yield per unit area of winter wheat and summer maize in four cities in 2020 was controlled within the range of 14.02 to 696.66 kg/hectare. It shows that the model can more accurately predict the yield per unit area of crops. The planning model for the benefit of irrigation water allocation generated three scenarios of rainfall level and four planting intentions, and compared the planned scenarios with the actual production benefits of the two crops in 2020. In a dry year, the possibility of planting areas for winter wheat and summer corn is optimized. Compared with the traditional deterministic planning method, the model takes into account the uncertain parameters, which helps decision makers seek better solutions under uncertain conditions.

Mapping Winter Wheat with Optical and SAR Images Based on Google Earth Engine in Henan Province, China

The timely and accurate acquisition of winter wheat acreage is crucial for food security. This study investigated the feasibility of extracting the spatial distribution map of winter wheat in Henan Province by using synthetic aperture radar (SAR, Sentinel-1A) and optical (Sentinel-2) images. Firstly, the SAR images were aggregated based on the growth period of winter wheat, and the optical images were aggregated based on the moderate resolution imaging spectroradiometer normalized difference vegetation index (MODIS-NDVI) curve. Then, five spectral features, two polarization features, and four texture features were selected as feature variables. Finally, the Google Earth Engine (GEE) cloud platform was employed to extract winter wheat acreage through the random forest (RF) algorithm. The results show that: (1) aggregated images based on the growth period of winter wheat and sensor characteristics can improve the mapping accuracy and efficiency; (2) the extraction accuracy of using only SAR images was improved with the accumulation of growth period. The extraction accuracy of using the SAR images in the full growth period reached 80.1%; and (3) the identification effect of integrated images was relatively good, which makes up for the shortcomings of SAR and optical images and improves the extraction accuracy of winter wheat.

Predicting rice yield at pixel scale through synthetic use of crop and deep learning models with satellite data in South and North Korea

Prediction of rice yields at pixel scale rather than county scale can benefit crop management and scientific understanding because it is useful for monitoring how crop yields respond to various agricultural systems and environmental factors. In this study, we propose a methodology for the early prediction of rice yield at pixel scale combining a crop model and a deep learning model for different agricultural systems throughout South and North Korea. Initially, satellite-integrated crop models were applied to obtain a pixel-scale reference rice yield. Then, the pixel-scale reference rice yields were used as target labels in the deep learning model to leverage the advantages of crop models. Models of five different deep learning network architectures were employed to help determine the hybrid structure of long-short term memory (LSTM) and one-dimensional convolutional neural network (1D-CNN) layers by predicting the optimal model about two months ahead of harvest time. The suggested model showed good performance [R² = 0.859, Nash-Sutcliffe model efficiency = 0.858, root mean squared error = 0.605 Mg ha^-1], with specific spatial patterns of rice yields for South and North Korea. Analysis of the relative importance of the input variables showed the water-related index and maximum temperature in North Korea and the vegetation indices and geographic variables in South Korea to be crucial for predicting rice yields. The proposed approach successfully predicted and diagnosed rice yield at the pixel scale for inaccessible locations where reliable ground measurements are not available, especially North Korea.

Wheat-Net: An Automatic Dense Wheat Spike Segmentation Method Based on an Optimized Hybrid Task Cascade Model

Precise segmentation of wheat spikes from a complex background is necessary for obtaining image-based phenotypic information of wheat traits such as yield estimation and spike morphology. A new instance segmentation method based on a Hybrid Task Cascade model was proposed to solve the wheat spike detection problem with improved detection results. In this study, wheat images were collected from fields where the environment varied both spatially and temporally. Res2Net50 was adopted as a backbone network, combined with multi-scale training, deformable convolutional networks, and Generic ROI Extractor for rich feature learning. The proposed methods were trained and validated, and the average precision (AP) obtained for the bounding box and mask was 0.904 and 0.907, respectively, and the accuracy for wheat spike counting was 99.29%. Comprehensive empirical analyses revealed that our method (Wheat-Net) performed well on challenging field-based datasets with mixed qualities, particularly those with various backgrounds and wheat spike adjacence/occlusion. These results provide evidence for dense wheat spike detection capabilities with masking, which is useful for not only wheat yield estimation but also spike morphology assessments.

DeepCap-Kcr: accurate identification and investigation of protein lysine crotonylation sites based on capsule network

Lysine crotonylation (Kcr) is a posttranslational modification widely detected in histone and nonhistone proteins. It plays a vital role in human disease progression and various cellular processes, including cell cycle, cell organization, chromatin remodeling and a key mechanism to increase proteomic diversity. Thus, accurate information on such sites is beneficial for both drug development and basic research. Existing computational methods can be improved to more effectively identify Kcr sites in proteins. In this study, we proposed a deep learning model, DeepCap-Kcr, a capsule network (CapsNet) based on a convolutional neural network (CNN) and long short-term memory (LSTM) for robust prediction of Kcr sites on histone and nonhistone proteins (mammals). The proposed model outperformed the existing CNN architecture Deep-Kcr and other well-established tools in most cases and provided promising outcomes for practical use; in particular, the proposed model characterized the internal hierarchical representation as well as the important features from multiple levels of abstraction automatically learned from a small number of samples. The trained model was well generalized in other species (papaya). Moreover, we showed the features and properties generated by the internal capsule layer that can explore the internal data distribution related to biological significance (as a motif detector). The source code and data are freely available at https://github.com/Jhabindra-bioinfo/DeepCap-Kcr.

Predicting Days to Maturity, Plant Height, and Grain Yield in Soybean: A Machine and Deep Learning Approach Using Multispectral Data

In soybean, there is a lack of research aiming to compare the performance of machine learning (ML) and deep learning (DL) methods to predict more than one agronomic variable, such as days to maturity (DM), plant height (PH), and grain yield (GY). As these variables are important to developing an overall precision farming model, we propose a machine learning approach to predict DM, PH, and GY for soybean cultivars based on multispectral bands. The field experiment considered 524 genotypes of soybeans in the 2017/2018 and 2018/2019 growing seasons and a multitemporal–multispectral dataset collected by embedded sensor in an unmanned aerial vehicle (UAV). We proposed a multilayer deep learning regression network, trained during 2000 epochs using an adaptive subgradient method, a random Gaussian initialization, and a 50% dropout in the first hidden layer for regularization. Three different scenarios, including only spectral bands, only vegetation indices, and spectral bands plus vegetation indices, were adopted to infer each variable (PH, DM, and GY). The DL model performance was compared against shallow learning methods such as random forest (RF), support vector machine (SVM), and linear regression (LR). The results indicate that our approach has the potential to predict soybean-related variables using multispectral bands only. Both DL and RF models presented a strong (r surpassing 0.77) prediction capacity for the PH variable, regardless of the adopted input variables group. Our results demonstrated that the DL model (r = 0.66) was superior to predict DM when the input variable was the spectral bands. For GY, all machine learning models evaluated presented similar performance (r ranging from 0.42 to 0.44) for each tested scenario. In conclusion, this study demonstrated an efficient approach to a computational solution capable of predicting multiple important soybean crop variables based on remote sensing data. Future research could benefit from the information presented here and be implemented in subsequent processes related to soybean cultivars or other types of agronomic crops.

A Comparative Analysis of Machine Learning Models for the Prediction of Insurance Uptake in Kenya

The role of insurance in financial inclusion and economic growth, in general, is immense and is increasingly being recognized. However, low uptake impedes the growth of the sector, hence the need for a model that robustly predicts insurance uptake among potential clients. This study undertook a two phase comparison of machine learning classifiers. Phase I had eight machine learning models compared for their performance in predicting the insurance uptake using 2016 Kenya FinAccessHousehold Survey data. Taking Phase I as a base in Phase II, random forest and XGBoost were compared with four deep learning classifiers using 2019 Kenya FinAccess Household Survey data. The random forest model trained on oversampled data showed the highest F1-score, accuracy, and precision. The area under the receiver operating characteristic curve was furthermore highest for random forest; hence, it could be construed as the most robust model for predicting the insurance uptake. Finally, the most important features in predicting insurance uptake as extracted from the random forest model were income, bank usage, and ability and willingness to support others. Hence, there is a need for a design and distribution of low income based products, and bancassurance could be said to be a plausible channel for the distribution of insurance products.

Maize Yield Prediction at an Early Developmental Stage Using Multispectral Images and Genotype Data for Preliminary Hybrid Selection

Assessing crop production in the field often requires breeders to wait until the end of the season to collect yield-related measurements, limiting the pace of the breeding cycle. Early prediction of crop performance can reduce this constraint by allowing breeders more time to focus on the highest-performing varieties. Here, we present a multimodal deep learning model for predicting the performance of maize (Zea mays) at an early developmental stage, offering the potential to accelerate crop breeding. We employed multispectral images and eight vegetation indices, collected by an uncrewed aerial vehicle approximately 60 days after sowing, over three consecutive growing cycles (2017, 2018 and 2019). The multimodal deep learning approach was used to integrate field management and genotype information with the multispectral data, providing context to the conditions that the plants experienced during the trial. Model performance was assessed using holdout data, in which the model accurately predicted the yield (RMSE 1.07 t/ha, a relative RMSE of 7.60% of 16 t/ha, and R2 score 0.73) and identified the majority of high-yielding varieties, outperforming previously published models for early yield prediction. The inclusion of vegetation indices was important for model performance, with a normalized difference vegetation index and green with normalized difference vegetation index contributing the most to model performance. The model provides a decision support tool, identifying promising lines early in the field trial.

Characterizing and forecasting the responses of tropical forest leaf phenology to El Nino by machine learning algorithms

Climate change and global warming have serious adverse impacts on tropical forests. In particular, climate change may induce changes in leaf phenology. However, in tropical dry forests where tree diversity is high, species responses to climate change differ. The objective of this research is to analyze the impact of climate variability on the leaf phenology in Thailand's tropical forests. Machine learning approaches were applied to model how leaf phenology in dry dipterocarp forest in Thailand responds to climate variability and El Niño. First, we used a Self-Organizing Map (SOM) to cluster mature leaf phenology at the species level. Then, leaf phenology patterns in each group along with litterfall phenology and climate data were analyzed according to their response time. After that, a Long Short-Term Memory neural network (LSTM) was used to create model to predict leaf phenology in dry dipterocarp forest. The SOM-based clustering was able to classify 92.24% of the individual trees. The result of mapping the clustering data with lag time analysis revealed that each cluster has a different lag time depending on the timing and amount of rainfall. Incorporating the time lags improved the performance of the litterfall prediction model, reducing the average root mean square percent error (RMSPE) from 14.35% to 12.06%. This study should help researchers understand how each species responds to climate change. The litterfall prediction model will be useful for managing dry dipterocarp forest especially with regards to forest fires.

Deep Learning Based Prediction on Greenhouse Crop Yield Combined TCN and RNN

Currently, greenhouses are widely applied for plant growth, and environmental parameters can also be controlled in the modern greenhouse to guarantee the maximum crop yield. In order to optimally control greenhouses' environmental parameters, one indispensable requirement is to accurately predict crop yields based on given environmental parameter settings. In addition, crop yield forecasting in greenhouses plays an important role in greenhouse farming planning and management, which allows cultivators and farmers to utilize the yield prediction results to make knowledgeable management and financial decisions. It is thus important to accurately predict the crop yield in a greenhouse considering the benefits that can be brought by accurate greenhouse crop yield prediction. In this work, we have developed a new greenhouse crop yield prediction technique, by combining two state-of-the-arts networks for temporal sequence processing-temporal convolutional network (TCN) and recurrent neural network (RNN). Comprehensive evaluations of the proposed algorithm have been made on multiple datasets obtained from multiple real greenhouse sites for tomato growing. Based on a statistical analysis of the root mean square errors (RMSEs) between the predicted and actual crop yields, it is shown that the proposed approach achieves more accurate yield prediction performance than both traditional machine learning methods and other classical deep neural networks. Moreover, the experimental study also shows that the historical yield information is the most important factor for accurately predicting future crop yields.

Machine Learning in Agriculture: A Comprehensive Updated Review

The digital transformation of agriculture has evolved various aspects of management into artificial intelligent systems for the sake of making value from the ever-increasing data originated from numerous sources. A subset of artificial intelligence, namely machine learning, has a considerable potential to handle numerous challenges in the establishment of knowledge-based farming systems. The present study aims at shedding light on machine learning in agriculture by thoroughly reviewing the recent scholarly literature based on keywords' combinations of "machine learning" along with "crop management", "water management", "soil management", and "livestock management", and in accordance with PRISMA guidelines. Only journal papers were considered eligible that were published within 2018-2020. The results indicated that this topic pertains to different disciplines that favour convergence research at the international level. Furthermore, crop management was observed to be at the centre of attention. A plethora of machine learning algorithms were used, with those belonging to Artificial Neural Networks being more efficient. In addition, maize and wheat as well as cattle and sheep were the most investigated crops and animals, respectively. Finally, a variety of sensors, attached on satellites and unmanned ground and aerial vehicles, have been utilized as a means of getting reliable input data for the data analyses. It is anticipated that this study will constitute a beneficial guide to all stakeholders towards enhancing awareness of the potential advantages of using machine learning in agriculture and contributing to a more systematic research on this topic.

Rice-Yield Prediction with Multi-Temporal Sentinel-2 Data and 3D CNN: A Case Study in Nepal

Crop yield estimation is a major issue of crop monitoring which remains particularly challenging in developing countries due to the problem of timely and adequate data availability. Whereas traditional agricultural systems mainly rely on scarce ground-survey data, freely available multi-temporal and multi-spectral remote sensing images are excellent tools to support these vulnerable systems by accurately monitoring and estimating crop yields before harvest. In this context, we introduce the use of Sentinel-2 (S2) imagery, with a medium spatial, spectral and temporal resolutions, to estimate rice crop yields in Nepal as a case study. Firstly, we build a new large-scale rice crop database (RicePAL) composed by multi-temporal S2 and climate/soil data from the Terai districts of Nepal. Secondly, we propose a novel 3D Convolutional Neural Network (CNN) adapted to these intrinsic data constraints for the accurate rice crop yield estimation. Thirdly, we study the effect of considering different temporal, climate and soil data configurations in terms of the performance achieved by the proposed approach and several state-of-the-art regression and CNN-based yield estimation methods. The extensive experiments conducted in this work demonstrate the suitability of the proposed CNN-based framework for rice crop yield estimation in the developing country of Nepal using S2 data.

Using satellite imagery to understand and promote sustainable development

Accurate and comprehensive measurements of a range of sustainable development outcomes are fundamental inputs into both research and policy. We synthesize the growing literature that uses satellite imagery to understand these outcomes, with a focus on approaches that combine imagery with machine learning. We quantify the paucity of ground data on key human-related outcomes and the growing abundance and improving resolution (spatial, temporal, and spectral) of satellite imagery. We then review recent machine learning approaches to model-building in the context of scarce and noisy training data, highlighting how this noise often leads to incorrect assessment of model performance. We quantify recent model performance across multiple sustainable development domains, discuss research and policy applications, explore constraints to future progress, and highlight research directions for the field.

Prediction of Crop Yield Using Phenological Information Extracted from Remote Sensing Vegetation Index

Phenology is an indicator of crop growth conditions, and is correlated with crop yields. In this study, a phenological approach based on a remote sensing vegetation index was explored to predict the yield in 314 counties within the US Corn Belt, divided into semi-arid and non-semi-arid regions. The Moderate Resolution Imaging Spectroradiometer (MODIS) data product MOD09Q1 was used to calculate the normalized difference vegetation index (NDVI) time series. According to the NDVI time series, we divided the corn growing season into four growth phases, calculated phenological information metrics (duration and rate) for each growth phase, and obtained the maximum correlation NDVI (Max-R2). Duration and rate represent crop growth days and rate, respectively. Max-R2 is the NDVI value with the most significant correlation with corn yield in the NDVI time series. We built three groups of yield regression models, including univariate models using phenological metrics and Max-R2, and multivariate models using phenological metrics, and multivariate models using phenological metrics combined with Max-R2 in the whole, semi-arid, and non-semi-arid regions, respectively, and compared the performance of these models. The results show that most phenological metrics had a statistically significant (p < 0.05) relationship with corn yield (maximum R2 = 0.44). Models established with phenological metrics realized yield prediction before harvest in the three regions with R2 = 0.64, 0.67, and 0.72. Compared with the univariate Max-R2 models, the accuracy of models built with Max-R2 and phenology metrics improved. Thus, the phenology metrics obtained from MODIS-NDVI accurately reflect the corn characteristics and can be used for large-scale yield prediction. Overall, this study showed that phenology metrics derived from remote sensing vegetation indexes could be used as crop yield prediction variables and provide a reference for data organization and yield prediction with physical crop significance.

Research on Runoff Simulations Using Deep-Learning Methods

Runoff simulations are of great significance to the planning management of water resources. Here, we discussed the influence of the model component, model parameters and model input on runoff modeling, taking Hanjiang River Basin as the research area. Convolution kernel and attention mechanism were introduced into an LSTM network, and a new data-driven model Conv-TALSTM was developed. The model parameters were analyzed based on the Conv-TALSTM, and the results suggested that the optimal parameters were greatly affected by the correlation between the input data and output data. We compared the performance of Conv-TALSTM and variant models (TALSTM, Conv-LSTM, LSTM), and found that Conv-TALSTM can reproduce high flow more accurately. Moreover, the results were comparable when the model was trained with meteorological or hydrological variables, whereas the peak values with hydrological data were closer to the observations. When the two datasets were combined, the performance of the model was better. Additionally, Conv-TALSTM was also compared with an ANN (artificial neural network) and Wetspa (a distributed model for Water and Energy Transfer between Soil, Plants and Atmosphere), which verified the advantages of Conv-TALSTM in peak simulations. This study provides a direction for improving the accuracy, simplifying model structure and shortening calculation time in runoff simulations.

Crop Yield Prediction Using Multitemporal UAV Data and Spatio-Temporal Deep Learning Models

Unmanned aerial vehicle (UAV) based remote sensing is gaining momentum worldwide in a variety of agricultural and environmental monitoring and modelling applications. At the same time, the increasing availability of yield monitoring devices in harvesters enables input-target mapping of in-season RGB and crop yield data in a resolution otherwise unattainable by openly availabe satellite sensor systems. Using time series UAV RGB and weather data collected from nine crop fields in Pori, Finland, we evaluated the feasibility of spatio-temporal deep learning architectures in crop yield time series modelling and prediction with RGB time series data. Using Convolutional Neural Networks (CNN) and Long-Short Term Memory (LSTM) networks as spatial and temporal base architectures, we developed and trained CNN-LSTM, convolutional LSTM and 3D-CNN architectures with full 15 week image frame sequences from the whole growing season of 2018. The best performing architecture, the 3D-CNN, was then evaluated with several shorter frame sequence configurations from the beginning of the season. With 3D-CNN, we were able to achieve 218.9 kg/ha mean absolute error (MAE) and 5.51% mean absolute percentage error (MAPE) performance with full length sequences. The best shorter length sequence performance with the same model was 292.8 kg/ha MAE and 7.17% MAPE with four weekly frames from the beginning of the season.

High-Resolution Soybean Yield Mapping Across the US Midwest Using Subfield Harvester Data

Cloud computing and freely available, high-resolution satellite data have enabled recent progress in crop yield mapping at fine scales. However, extensive validation data at a matching resolution remain uncommon or infeasible due to data availability. This has limited the ability to evaluate different yield estimation models and improve understanding of key features useful for yield estimation in both data-rich and data-poor contexts. Here, we assess machine learning models’ capacity for soybean yield prediction using a unique ground-truth dataset of high-resolution (5 m) yield maps generated from combine harvester yield monitor data for over a million field-year observations across the Midwestern United States from 2008 to 2018. First, we compare random forest (RF) implementations, testing a range of feature engineering approaches using Sentinel-2 and Landsat spectral data for 20- and 30-m scale yield prediction. We find that Sentinel-2-based models can explain up to 45% of out-of-sample yield variability from 2017 to 2018 (r2 = 0.45), while Landsat models explain up to 43% across the longer 2008–2018 period. Using discrete Fourier transforms, or harmonic regressions, to capture soybean phenology improved the Landsat-based model considerably. Second, we compare RF models trained using this ground-truth data to models trained on available county-level statistics. We find that county-level models rely more heavily on just a few predictors, namely August weather covariates (vapor pressure deficit, rainfall, temperature) and July and August near-infrared observations. As a result, county-scale models perform relatively poorly on field-scale validation (r2 = 0.32), especially for high-yielding fields, but perform similarly to field-scale models when evaluated at the county scale (r2 = 0.82). Finally, we test whether our findings on variable importance can inform a simple, generalizable framework for regions or time periods beyond ground data availability. To do so, we test improvements to a Scalable Crop Yield Mapper (SCYM) approach that uses crop simulations to train statistical models for yield estimation. Based on findings from our RF models, we employ harmonic regressions to estimate peak vegetation index (VI) and a VI observation 30 days later, with August rainfall as the sole weather covariate in our new SCYM model. Modifications improved SCYM’s explained variance (r2 = 0.27 at the 30 m scale) and provide a new, parsimonious model.

Leaf Nitrogen Concentration and Plant Height Prediction for Maize Using UAV-Based Multispectral Imagery and Machine Learning Techniques

Under ideal conditions of nitrogen (N), maize (Zea mays L.) can grow to its full potential, reaching maximum plant height (PH). As a rapid and nondestructive approach, the analysis of unmanned aerial vehicles (UAV)-based imagery may be of assistance to estimate N and height. The main objective of this study is to present an approach to predict leaf nitrogen concentration (LNC, g kg−1) and PH (m) with machine learning techniques and UAV-based multispectral imagery in maize plants. An experiment with 11 maize cultivars under two rates of N fertilization was carried during the 2017/2018 and 2018/2019 crop seasons. The spectral vegetation indices (VI) normalized difference vegetation index (NDVI), normalized difference red-edge index (NDRE), green normalized difference vegetation (GNDVI), and the soil adjusted vegetation index (SAVI) were extracted from the images and, in a computational system, used alongside the spectral bands as input parameters for different machine learning models. A randomized 10-fold cross-validation strategy, with a total of 100 replicates, was used to evaluate the performance of 9 supervised machine learning (ML) models using the Pearson’s correlation coefficient (r), mean absolute error (MAE), coefficient of regression (R²), and root mean square error (RMSE) metrics. The results indicated that the random forest (RF) algorithm performed better, with r and RMSE, respectively, of 0.91 and 1.9 g.kg−¹ for LNC, and 0.86 and 0.17 m for PH. It was also demonstrated that VIs contributed more to the algorithm’s performances than individual spectral bands. This study concludes that the RF model is appropriate to predict both agronomic variables in maize and may help farmers to monitor their plants based upon their LNC and PH diagnosis and use this knowledge to improve their production rates in the subsequent seasons.