Machine Learning in Agriculture: A Review

Weed and Corn Seedling Detection in Field Based on Multi Feature Fusion and Support Vector Machine

Detection of weeds and crops is the key step for precision spraying using the spraying herbicide robot and precise fertilization for the agriculture machine in the field. On the basis of k-mean clustering image segmentation using color information and connected region analysis, a method combining multi feature fusion and support vector machine (SVM) was proposed to identify and detect the position of corn seedlings and weeds, to reduce the harm of weeds on corn growth, and to achieve accurate fertilization, thereby realizing precise weeding or fertilizing. First, the image dataset for weed and corn seedling classification in the corn seedling stage was established. Second, many different features of corn seedlings and weeds were extracted, and dimensionality was reduced by principal component analysis, including the histogram of oriented gradient feature, rotation invariant local binary pattern (LBP) feature, Hu invariant moment feature, Gabor feature, gray level co-occurrence matrix, and gray level-gradient co-occurrence matrix. Then, the classifier training based on SVM was conducted to obtain the recognition model for corn seedlings and weeds. The comprehensive recognition performance of single feature or different fusion strategies for six features is compared and analyzed, and the optimal feature fusion strategy is obtained. Finally, by utilizing the actual corn seedling field images, the proposed weed and corn seedling detection method effect was tested. LAB color space and K-means clustering were used to achieve image segmentation. Connected component analysis was adopted to remove small objects. The previously trained recognition model was utilized to identify and label each connected region to identify and detect weeds and corn seedlings. The experimental results showed that the fusion feature combination of rotation invariant LBP feature and gray level-gradient co-occurrence matrix based on SVM classifier obtained the highest classification accuracy and accurately detected all kinds of weeds and corn seedlings. It provided information on weed and crop positions to the spraying herbicide robot for accurate spraying or to the precise fertilization machine for accurate fertilizing.

Thermal Imaging for Plant Stress Detection and Phenotyping

In the last few years, large efforts have been made to develop new methods to optimize stress detection in crop fields. Thus, plant phenotyping based on imaging techniques has become an essential tool in agriculture. In particular, leaf temperature is a valuable indicator of the physiological status of plants, responding to both biotic and abiotic stressors. Often combined with other imaging sensors and data-mining techniques, thermography is crucial in the implementation of a more automatized, precise and sustainable agriculture. However, thermal data need some corrections related to the environmental and measuring conditions in order to achieve a correct interpretation of the data. This review focuses on the state of the art of thermography applied to the detection of biotic stress. The work will also revise the most important abiotic stress factors affecting the measurements as well as practical issues that need to be considered in order to implement this technique, particularly at the field scale.

An embedded system for the automated generation of labeled plant images to enable machine learning applications in agriculture

A lack of sufficient training data, both in terms of variety and quantity, is often the bottleneck in the development of machine learning (ML) applications in any domain. For agricultural applications, ML-based models designed to perform tasks such as autonomous plant classification will typically be coupled to just one or perhaps a few plant species. As a consequence, each crop-specific task is very likely to require its own specialized training data, and the question of how to serve this need for data now often overshadows the more routine exercise of actually training such models. To tackle this problem, we have developed an embedded robotic system to automatically generate and label large datasets of plant images for ML applications in agriculture. The system can image plants from virtually any angle, thereby ensuring a wide variety of data; and with an imaging rate of up to one image per second, it can produce lableled datasets on the scale of thousands to tens of thousands of images per day. As such, this system offers an important alternative to time- and cost-intensive methods of manual generation and labeling. Furthermore, the use of a uniform background made of blue keying fabric enables additional image processing techniques such as background replacement and image segementation. It also helps in the training process, essentially forcing the model to focus on the plant features and eliminating random correlations. To demonstrate the capabilities of our system, we generated a dataset of over 34,000 labeled images, with which we trained an ML-model to distinguish grasses from non-grasses in test data from a variety of sources. We now plan to generate much larger datasets of Canadian crop plants and weeds that will be made publicly available in the hope of further enabling ML applications in the agriculture sector.

Design and Development of a Smart Variable Rate Sprayer Using Deep Learning

The uniform application (UA) of agrochemicals results in the over-application of harmful chemicals, increases crop input costs, and deteriorates the environment when compared with variable rate application (VA). A smart variable rate sprayer (SVRS) was designed, developed, and tested using deep learning (DL) for VA application of agrochemicals. Real-time testing of the SVRS took place for detecting and spraying and/or skipping lambsquarters weed and early blight infected and healthy potato plants. About 24,000 images were collected from potato fields in Prince Edward Island and New Brunswick under varying sunny, cloudy, and partly cloudy conditions and processed/trained using YOLOv3 and tiny-YOLOv3 models. Due to faster performance, the tiny-YOLOv3 was chosen to deploy in SVRS. A laboratory experiment was designed under factorial arrangements, where the two spraying techniques (UA and VA) and the three weather conditions (cloudy, partly cloudy, and sunny) were the two independent variables with spray volume consumption as a response variable. The experimental treatments had six repetitions in a 2 × 3 factorial design. Results of the two-way ANOVA showed a significant effect of spraying application techniques on volume consumption of spraying liquid (p-value < 0.05). There was no significant effect of weather conditions and interactions between the two independent variables on volume consumption during weeds and simulated diseased plant detection experiments (p-value > 0.05). The SVRS was able to save 42 and 43% spraying liquid during weeds and simulated diseased plant detection experiments, respectively. Water sensitive papers’ analysis showed the applicability of SVRS for VA with >40% savings of spraying liquid by SVRS when compared with UA. Field applications of this technique would reduce the crop input costs and the environmental risks in conditions (weed and disease) like experimental testing.

Machine Learning Techniques for Soybean Charcoal Rot Disease Prediction

Early prediction of pathogen infestation is a key factor to reduce the disease spread in plants. Macrophomina phaseolina (Tassi) Goid, as one of the main causes of charcoal rot disease, suppresses the plant productivity significantly. Charcoal rot disease is one of the most severe threats to soybean productivity. Prediction of this disease in soybeans is very tedious and non-practical using traditional approaches. Machine learning (ML) techniques have recently gained substantial traction across numerous domains. ML methods can be applied to detect plant diseases, prior to the full appearance of symptoms. In this paper, several ML techniques were developed and examined for prediction of charcoal rot disease in soybean for a cohort of 2,000 healthy and infected plants. A hybrid set of physiological and morphological features were suggested as inputs to the ML models. All developed ML models were performed better than 90% in terms of accuracy. Gradient Tree Boosting (GBT) was the best performing classifier which obtained 96.25% and 97.33% in terms of sensitivity and specificity. Our findings supported the applicability of ML especially GBT for charcoal rot disease prediction in a real environment. Moreover, our analysis demonstrated the importance of including physiological featured in the learning. The collected dataset and source code can be found in https://github.com/Elham-khalili/Soybean-Charcoal-Rot-Disease-Prediction-Dataset-code.

Proximal Methods for Plant Stress Detection Using Optical Sensors and Machine Learning

Plant stresses have been monitored using the imaging or spectrometry of plant leaves in the visible (red-green-blue or RGB), near-infrared (NIR), infrared (IR), and ultraviolet (UV) wavebands, often augmented by fluorescence imaging or fluorescence spectrometry. Imaging at multiple specific wavelengths (multi-spectral imaging) or across a wide range of wavelengths (hyperspectral imaging) can provide exceptional information on plant stress and subsequent diseases. Digital cameras, thermal cameras, and optical filters have become available at a low cost in recent years, while hyperspectral cameras have become increasingly more compact and portable. Furthermore, smartphone cameras have dramatically improved in quality, making them a viable option for rapid, on-site stress detection. Due to these developments in imaging technology, plant stresses can be monitored more easily using handheld and field-deployable methods. Recent advances in machine learning algorithms have allowed for images and spectra to be analyzed and classified in a fully automated and reproducible manner, without the need for complicated image or spectrum analysis methods. This review will highlight recent advances in portable (including smartphone-based) detection methods for biotic and abiotic stresses, discuss data processing and machine learning techniques that can produce results for stress identification and classification, and suggest future directions towards the successful translation of these methods into practical use.

Groundwater Prediction Using Machine-Learning Tools

Predicting groundwater availability is important to water sustainability and drought mitigation. Machine-learning tools have the potential to improve groundwater prediction, thus enabling resource planners to: (1) anticipate water quality in unsampled areas or depth zones; (2) design targeted monitoring programs; (3) inform groundwater protection strategies; and (4) evaluate the sustainability of groundwater sources of drinking water. This paper proposes a machine-learning approach to groundwater prediction with the following characteristics: (i) the use of a regression-based approach to predict full groundwater images based on sequences of monthly groundwater maps; (ii) strategic automatic feature selection (both local and global features) using extreme gradient boosting; and (iii) the use of a multiplicity of machine-learning techniques (extreme gradient boosting, multivariate linear regression, random forests, multilayer perceptron and support vector regression). Of these techniques, support vector regression consistently performed best in terms of minimizing root mean square error and mean absolute error. Furthermore, including a global feature obtained from a Gaussian Mixture Model produced models with lower error than the best which could be obtained with local geographical features.

Machine learning-enabled multiplexed microfluidic sensors

High-throughput, cost-effective, and portable devices can enhance the performance of point-of-care tests. Such devices are able to acquire images from samples at a high rate in combination with microfluidic chips in point-of-care applications. However, interpreting and analyzing the large amount of acquired data is not only a labor-intensive and time-consuming process, but also prone to the bias of the user and low accuracy. Integrating machine learning (ML) with the image acquisition capability of smartphones as well as increasing computing power could address the need for high-throughput, accurate, and automatized detection, data processing, and quantification of results. Here, ML-supported diagnostic technologies are presented. These technologies include quantification of colorimetric tests, classification of biological samples (cells and sperms), soft sensors, assay type detection, and recognition of the fluid properties. Challenges regarding the implementation of ML methods, including the required number of data points, image acquisition prerequisites, and execution of data-limited experiments are also discussed.

A Machine Learning Approach in Analyzing Bioaccumulation of Heavy Metals in Turbot Tissues

Metals are considered to be one of the most hazardous substances due to their potential for accumulation, magnification, persistence, and wide distribution in water, sediments, and aquatic organisms. Demersal fish species, such as turbot (Psetta maxima maeotica), are accepted by the scientific communities as suitable bioindicators of heavy metal pollution in the aquatic environment. The present study uses a machine learning approach, which is based on multiple linear and non-linear models, in order to effectively estimate the concentrations of heavy metals in both turbot muscle and liver tissues. For multiple linear regression (MLR) models, the stepwise method was used, while non-linear models were developed by applying random forest (RF) algorithm. The models were based on data that were provided from scientific literature, attributed to 11 heavy metals (As, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, Ni, Zn) from both muscle and liver tissues of turbot exemplars. Significant MLR models were recorded for Ca, Fe, Mg, and Na in muscle tissue and K, Cu, Zn, and Na in turbot liver tissue. The non-linear tree-based RF prediction models (over 70% prediction accuracy) were identified for As, Cd, Cu, K, Mg, and Zn in muscle tissue and As, Ca, Cd, Mg, and Fe in turbot liver tissue. Both machine learning MLR and non-linear tree-based RF prediction models were identified to be suitable for predicting the heavy metal concentration from both turbot muscle and liver tissues. The models can be used for improving the knowledge and economic efficiency of linked heavy metals food safety and environment pollution studies.

ISOBlue HD: An Open-Source Platform for Collecting Context-Rich Agricultural Machinery Datasets

This paper introduces an open-source platform called ISOBlue HD for acquisition of context-rich data from agricultural machinery. We call these datasets context-rich, because they enable the identification of machine status and farming logistics by properly labeling, fusing, and processing the data. The system includes a single board computer, a cellular modem, local storage, and power-over-ethernet switch to sensors. The system allows remote diagnostics and access, automatic startup/shut down with vehicle operations, and uses Apache Kafka to enable robust data exchange. ISOBlue HD was deployed in a combine harvester during a 2019 wheat harvest for simultaneously capturing 69 million CAN frames, 230,000 GPS points, and 437 GB of video data, focusing on header status and operator actions over 84 h of harvest time. Analyses of the collected data demonstrate that contextual knowledge can be inferred on harvest logistics (paths, speeds, header status, material transfer) and sensor data semantics.

The potential of remote sensing and artificial intelligence as tools to improve the resilience of agriculture production systems

Modern agriculture and food production systems are facing increasing pressures from climate change, land and water availability, and, more recently, a pandemic. These factors are threatening the environmental and economic sustainability of current and future food supply systems. Scientific and technological innovations are needed more than ever to secure enough food for a fast-growing global population. Scientific advances have led to a better understanding of how various components of the agricultural system interact, from the cell to the field level. Despite incredible advances in genetic tools over the past few decades, our ability to accurately assess crop status in the field, at scale, has been severely lacking until recently. Thanks to recent advances in remote sensing and Artificial Intelligence (AI), we can now quantify field scale phenotypic information accurately and integrate the big data into predictive and prescriptive management tools. This review focuses on the use of recent technological advances in remote sensing and AI to improve the resilience of agricultural systems, and we will present a unique opportunity for the development of prescriptive tools needed to address the next decade's agricultural and human nutrition challenges.

Plant Disease Classification: A Comparative Evaluation of Convolutional Neural Networks and Deep Learning Optimizers

Recently, plant disease classification has been done by various state-of-the-art deep learning (DL) architectures on the publicly available/author generated datasets. This research proposed the deep learning-based comparative evaluation for the classification of plant disease in two steps. Firstly, the best convolutional neural network (CNN) was obtained by conducting a comparative analysis among well-known CNN architectures along with modified and cascaded/hybrid versions of some of the DL models proposed in the recent researches. Secondly, the performance of the best-obtained model was attempted to improve by training through various deep learning optimizers. The comparison between various CNNs was based on performance metrics such as validation accuracy/loss, F1-score, and the required number of epochs. All the selected DL architectures were trained in the PlantVillage dataset which contains 26 different diseases belonging to 14 respective plant species. Keras with TensorFlow backend was used to train deep learning architectures. It is concluded that the Xception architecture trained with the Adam optimizer attained the highest validation accuracy and F1-score of 99.81% and 0.9978 respectively which is comparatively better than the previous approaches and it proves the novelty of the work. Therefore, the method proposed in this research can be applied to other agricultural applications for transparent detection and classification purposes.

Review of the State of the Art of Deep Learning for Plant Diseases: A Broad Analysis and Discussion

Deep learning (DL) represents the golden era in the machine learning (ML) domain, and it has gradually become the leading approach in many fields. It is currently playing a vital role in the early detection and classification of plant diseases. The use of ML techniques in this field is viewed as having brought considerable improvement in cultivation productivity sectors, particularly with the recent emergence of DL, which seems to have increased accuracy levels. Recently, many DL architectures have been implemented accompanying visualisation techniques that are essential for determining symptoms and classifying plant diseases. This review investigates and analyses the most recent methods, developed over three years leading up to 2020, for training, augmentation, feature fusion and extraction, recognising and counting crops, and detecting plant diseases, including how these methods can be harnessed to feed deep classifiers and their effects on classifier accuracy.

Analysis of Cattle Social Transitional Behaviour: Attraction and Repulsion

Understanding social interactions in livestock groups could improve management practices, but this can be difficult and time-consuming using traditional methods of live observations and video recordings. Sensor technologies and machine learning techniques could provide insight not previously possible. In this study, based on the animals' location information acquired by a new cooperative wireless localisation system, unsupervised machine learning approaches were performed to identify the social structure of a small group of cattle yearlings (n=10) and the social behaviour of an individual. The paper first defined the affinity between an animal pair based on the ranks of their distance. Unsupervised clustering algorithms were then performed, including K-means clustering and agglomerative hierarchical clustering. In particular, K-means clustering was applied based on logical and physical distance. By comparing the clustering result based on logical distance and physical distance, the leader animals and the influence of an individual in a herd of cattle were identified, which provides valuable information for studying the behaviour of animal herds. Improvements in device robustness and replication of this work would confirm the practical application of this technology and analysis methodologies.

Modified Red Blue Vegetation Index for Chlorophyll Estimation and Yield Prediction of Maize from Visible Images Captured by UAV

The vegetation index (VI) has been successfully used to monitor the growth and to predict the yield of agricultural crops. In this paper, a long-term observation was conducted for the yield prediction of maize using an unmanned aerial vehicle (UAV) and estimations of chlorophyll contents using SPAD-502. A new vegetation index termed as modified red blue VI (MRBVI) was developed to monitor the growth and to predict the yields of maize by establishing relationships between MRBVI- and SPAD-502-based chlorophyll contents. The coefficients of determination (R2s) were 0.462 and 0.570 in chlorophyll contents' estimations and yield predictions using MRBVI, and the results were relatively better than the results from the seven other commonly used VI approaches. All VIs during the different growth stages of maize were calculated and compared with the measured values of chlorophyll contents directly, and the relative error (RE) of MRBVI is the lowest at 0.355. Further, machine learning (ML) methods such as the backpropagation neural network model (BP), support vector machine (SVM), random forest (RF), and extreme learning machine (ELM) were adopted for predicting the yields of maize. All VIs calculated for each image captured during important phenological stages of maize were set as independent variables and the corresponding yields of each plot were defined as dependent variables. The ML models used the leave one out method (LOO), where the root mean square errors (RMSEs) were 2.157, 1.099, 1.146, and 1.698 (g/hundred grain weight) for BP, SVM, RF, and ELM. The mean absolute errors (MAEs) were 1.739, 0.886, 0.925, and 1.356 (g/hundred grain weight) for BP, SVM, RF, and ELM, respectively. Thus, the SVM method performed better in predicting the yields of maize than the other ML methods. Therefore, it is strongly suggested that the MRBVI calculated from images acquired at different growth stages integrated with advanced ML methods should be used for agricultural- and ecological-related chlorophyll estimation and yield predictions.

Maize Kernel Abortion Recognition and Classification Using Binary Classification Machine Learning Algorithms and Deep Convolutional Neural Networks

Maize kernel traits such as kernel length, kernel width, and kernel number determine the total kernel weight and, consequently, maize yield. Therefore, the measurement of kernel traits is important for maize breeding and the evaluation of maize yield. There are a few methods that allow the extraction of ear and kernel features through image processing. We evaluated the potential of deep convolutional neural networks and binary machine learning (ML) algorithms (logistic regression (LR), support vector machine (SVM), AdaBoost (ADB), Classification tree (CART), and the K-Neighbor (kNN)) for accurate maize kernel abortion detection and classification. The algorithms were trained using 75% of 66 total images, and the remaining 25% was used for testing their performance. Confusion matrix, classification accuracy, and precision were the major metrics in evaluating the performance of the algorithms. The SVM and LR algorithms were highly accurate and precise (100%) under all the abortion statuses, while the remaining algorithms had a performance greater than 95%. Deep convolutional neural networks were further evaluated using different activation and optimization techniques. The best performance (100% accuracy) was reached using the rectifier linear unit (ReLu) activation procedure and the Adam optimization technique. Maize ear with abortion were accurately detected by all tested algorithms with minimum training and testing time compared to ear without abortion. The findings suggest that deep convolutional neural networks can be used to detect the maize ear abortion status supplemented with the binary machine learning algorithms in maize breading programs. By using a convolution neural network (CNN) method, more data (big data) can be collected and processed for hundreds of maize ears, accelerating the phenotyping process.

A Deep Learning Approach for Weed Detection in Lettuce Crops Using Multispectral Images

Weed management is one of the most important aspects of crop productivity; knowing the amount and the locations of weeds has been a problem that experts have faced for several decades. This paper presents three methods for weed estimation based on deep learning image processing in lettuce crops, and we compared them to visual estimations by experts. One method is based on support vector machines (SVM) using histograms of oriented gradients (HOG) as feature descriptor. The second method was based in YOLOV3 (you only look once V3), taking advantage of its robust architecture for object detection, and the third one was based on Mask R-CNN (region based convolutional neural network) in order to get an instance segmentation for each individual. These methods were complemented with a NDVI index (normalized difference vegetation index) as a background subtractor for removing non photosynthetic objects. According to chosen metrics, the machine and deep learning methods had F1-scores of 88%, 94%, and 94% respectively, regarding to crop detection. Subsequently, detected crops were turned into a binary mask and mixed with the NDVI background subtractor in order to detect weed in an indirect way. Once the weed image was obtained, the coverage percentage of weed was calculated by classical image processing methods. Finally, these performances were compared with the estimations of a set from weed experts through a Bland–Altman plot, intraclass correlation coefficients (ICCs) and Dunn’s test to obtain statistical measurements between every estimation (machine-human); we found that these methods improve accuracy on weed coverage estimation and minimize subjectivity in human-estimated data.

Remote Sensing of Diseases

Detection, identification, and quantification of plant diseases by sensor techniques are expected to enable a more precise disease control, as sensors are sensitive, objective, and highly available for disease assessment. Recent progress in sensor technology and data processing is very promising; nevertheless, technical constraints and issues inherent to variability in host-pathogen interactions currently limit the use of sensors in various fields of application. The information from spectral [e.g., RGB (red, green, blue)], multispectral, and hyperspectral sensors that measure reflectance, fluorescence, and emission of radiation or from electronic noses that detect volatile organic compounds released from plants or pathogens, as well as the potential of sensors to characterize the health status of crops, is evaluated based on the recent literature. Phytopathological aspects of remote sensing of plant diseases across different scales and for various purposes are discussed, including spatial disease patterns, epidemic spread of pathogens, crop characteristics, and links to disease control. Future challenges in sensor use are identified.

Weber Texture Local Descriptor for Identification of Group-Housed Pigs

The individual identification of group-housed pigs plays an important role in breeding process management and individual behavior analysis. Recently, livestock identification methods based on the side view or face image have strict requirements on the position and posture of livestock, which poses a challenge for the application of the monitoring scene of group-housed pigs. To address the issue above, a Weber texture local descriptor (WTLD) is proposed for the identification of group-housed pigs by extracting the local features of back hair, skin texture, spots, and so on. By calculating the differential excitation and multi-directional information of pixels, the local structure features of the main direction are fused to enhance the description ability of features. The experimental results show that the proposed WTLD achieves higher recognition rates with a lower feature dimension. This method can identify pig individuals with different positions and postures in the pig house. Without limitations on pig movement, this method can facilitate the identification of individual pigs with greater convenience and universality.

A Review on Drone-Based Data Solutions for Cereal Crops

Food security is a longstanding global issue over the last few centuries. Eradicating hunger and all forms of malnutrition by 2030 is still a key challenge. The COVID-19 pandemic has placed additional stress on food production, demand, and supply chain systems; majorly impacting cereal crop producer and importer countries. Short food supply chain based on the production from local farms is less susceptible to travel and export bans and works as a smooth system in the face of these stresses. Local drone-based data solutions can provide an opportunity to address these challenges. This review aims to present a deeper understanding of how the drone-based data solutions can help to combat food insecurity caused due to the pandemic, zoonotic diseases, and other food shocks by enhancing cereal crop productivity of small-scale farming systems in low-income countries. More specifically, the review covers sensing capabilities, promising algorithms, and methods, and added-value of novel machine learning algorithms for local-scale monitoring, biomass and yield estimation, and mapping of them. Finally, we present the opportunities for linking information from citizen science, internet of things (IoT) based on low-cost sensors and drone-based information to satellite data for upscaling crop yield estimation to a larger geographical extent within the Earth Observation umbrella.

Autonomous Navigation of a Center-Articulated and Hydrostatic Transmission Rover using a Modified Pure Pursuit Algorithm in a Cotton Field

This study proposes an algorithm that controls an autonomous, multi-purpose, center-articulated hydrostatic transmission rover to navigate along crop rows. This multi-purpose rover (MPR) is being developed to harvest undefoliated cotton to expand the harvest window to up to 50 days. The rover would harvest cotton in teams by performing several passes as the bolls become ready to harvest. We propose that a small robot could make cotton production more profitable for farmers and more accessible to owners of smaller plots of land who cannot afford large tractors and harvesting equipment. The rover was localized with a low-cost Real-Time Kinematic Global Navigation Satellite System (RTK-GNSS), encoders, and Inertial Measurement Unit (IMU)s for heading. Robot Operating System (ROS)-based software was developed to harness the sensor information, localize the rover, and execute path following controls. To test the localization and modified pure-pursuit path-following controls, first, GNSS waypoints were obtained by manually steering the rover over the rows followed by the rover autonomously driving over the rows. The results showed that the robot achieved a mean absolute error (MAE) of 0.04 m, 0.06 m, and 0.09 m for the first, second and third passes of the experiment, respectively. The robot achieved an MAE of 0.06 m. When turning at the end of the row, the MAE from the RTK-GNSS-generated path was 0.24 m. The turning errors were acceptable for the open field at the end of the row. Errors while driving down the row did damage the plants by moving close to the plants' stems, and these errors likely would not impede operations designed for the MPR. Therefore, the designed rover and control algorithms are good and can be used for cotton harvesting operations.

Site-specific machine learning predictive fertilization models for potato crops in Eastern Canada

Statistical modeling is commonly used to relate the performance of potato (Solanum tuberosum L.) to fertilizer requirements. Prescribing optimal nutrient doses is challenging because of the involvement of many variables including weather, soils, land management, genotypes, and severity of pests and diseases. Where sufficient data are available, machine learning algorithms can be used to predict crop performance. The objective of this study was to determine an optimal model predicting nitrogen, phosphorus and potassium requirements for high tuber yield and quality (size and specific gravity) as impacted by weather, soils and land management variables. We exploited a data set of 273 field experiments conducted from 1979 to 2017 in Quebec (Canada). We developed, evaluated and compared predictions from a hierarchical Mitscherlich model, k-nearest neighbors, random forest, neural networks and Gaussian processes. Machine learning models returned R2 values of 0.49-0.59 for tuber marketable yield prediction, which were higher than the Mitscherlich model R2 (0.37). The models were more likely to predict medium-size tubers (R2 = 0.60-0.69) and tuber specific gravity (R2 = 0.58-0.67) than large-size tubers (R2 = 0.55-0.64) and marketable yield. Response surfaces from the Mitscherlich model, neural networks and Gaussian processes returned smooth responses that agreed more with actual evidence than discontinuous curves derived from k-nearest neighbors and random forest models. When conditioned to obtain optimal dosages from dose-response surfaces given constant weather, soil and land management conditions, some disagreements occurred between models. Due to their built-in ability to develop recommendations within a probabilistic risk-assessment framework, Gaussian processes stood out as the most promising algorithm to support decisions that minimize economic or agronomic risks.

Moving beyond the Technology: A Socio-technical Roadmap for Low-Cost Water Sensor Network Applications

In this paper, we critically review the current state-of-the-art for sensor network applications and approaches that have developed in response to the recent rise of low-cost technologies. We specifically focus on water-related low-cost sensor networks, and conceptualize them as socio-technical systems that can address resource management challenges and opportunities at three scales of resolution: (1) technologies, (2) users and scenarios, and (3) society and communities. Building this argument, first we identify a general structure for building low-cost sensor networks by assembling technical components across configuration levels. Second, we identify four application categories, namely operational monitoring, scientific research, system optimization, and community development, each of which has different technical and nontechnical configurations that determine how, where, by whom, and for what purpose low-cost sensor networks are used. Third, we discuss the governance factors (e.g., stakeholders and users, networks sustainability and maintenance, application scenarios, and integrated design) and emerging technical opportunities that we argue need to be considered to maximize the added value and long-term societal impact of the next generation of sensor network applications. We conclude that consideration of the full range of socio-technical issues is essential to realize the full potential of sensor network technologies for society and the environment.

Large-Scale Phenotyping of Livestock Welfare in Commercial Production Systems: A New Frontier in Animal Breeding

Genomic breeding programs have been paramount in improving the rates of genetic progress of productive efficiency traits in livestock. Such improvement has been accompanied by the intensification of production systems, use of a wider range of precision technologies in routine management practices, and high-throughput phenotyping. Simultaneously, a greater public awareness of animal welfare has influenced livestock producers to place more emphasis on welfare relative to production traits. Therefore, management practices and breeding technologies in livestock have been developed in recent years to enhance animal welfare. In particular, genomic selection can be used to improve livestock social behavior, resilience to disease and other stress factors, and ease habituation to production system changes. The main requirements for including novel behavioral and welfare traits in genomic breeding schemes are: (1) to identify traits that represent the biological mechanisms of the industry breeding goals; (2) the availability of individual phenotypic records measured on a large number of animals (ideally with genomic information); (3) the derived traits are heritable, biologically meaningful, repeatable, and (ideally) not highly correlated with other traits already included in the selection indexes; and (4) genomic information is available for a large number of individuals (or genetically close individuals) with phenotypic records. In this review, we (1) describe a potential route for development of novel welfare indicator traits (using ideal phenotypes) for both genetic and genomic selection schemes; (2) summarize key indicator variables of livestock behavior and welfare, including a detailed assessment of thermal stress in livestock; (3) describe the primary statistical and bioinformatic methods available for large-scale data analyses of animal welfare; and (4) identify major advancements, challenges, and opportunities to generate high-throughput and large-scale datasets to enable genetic and genomic selection for improved welfare in livestock. A wide variety of novel welfare indicator traits can be derived from information captured by modern technology such as sensors, automatic feeding systems, milking robots, activity monitors, video cameras, and indirect biomarkers at the cellular and physiological levels. The development of novel traits coupled with genomic selection schemes for improved welfare in livestock can be feasible and optimized based on recently developed (or developing) technologies. Efficient implementation of genetic and genomic selection for improved animal welfare also requires the integration of a multitude of scientific fields such as cell and molecular biology, neuroscience, immunology, stress physiology, computer science, engineering, quantitative genomics, and bioinformatics.

Learning from Data to Optimize Control in Precision Farming

Precision farming is one way of many to meet a 55 percent increase in global demand for agricultural products on current agricultural land by 2050 at reduced need of fertilizers and efficient use of water resources. The catalyst for the emergence of precision farming has been satellite positioning and navigation followed by Internet-of-Things, generating vast information that can be used to optimize farming processes in real-time. Statistical tools from data mining, predictive modeling, and machine learning analyze patterns in historical data, to make predictions about future events as well as intelligent actions. This special issue presents the latest development in statistical inference, machine learning, and optimum control for precision farming.

Predicting and Mapping of Soil Organic Carbon Using Machine Learning Algorithms in Northern Iran

Estimation of the soil organic carbon (SOC) content is of utmost importance in understanding the chemical, physical, and biological functions of the soil. This study proposes machine learning algorithms of support vector machines (SVM), artificial neural networks (ANN), regression tree, random forest (RF), extreme gradient boosting (XGBoost), and conventional deep neural network (DNN) for advancing prediction models of SOC. Models are trained with 1879 composite surface soil samples, and 105 auxiliary data as predictors. The genetic algorithm is used as a feature selection approach to identify effective variables. The results indicate that precipitation is the most important predictor driving 14.9% of SOC spatial variability followed by the normalized difference vegetation index (12.5%), day temperature index of moderate resolution imaging spectroradiometer (10.6%), multiresolution valley bottom flatness (8.7%) and land use (8.2%), respectively. Based on 10-fold cross-validation, the DNN model reported as a superior algorithm with the lowest prediction error and uncertainty. In terms of accuracy, DNN yielded a mean absolute error of 0.59%, a root mean squared error of 0.75%, a coefficient of determination of 0.65, and Lin’s concordance correlation coefficient of 0.83. The SOC content was the highest in udic soil moisture regime class with mean values of 3.71%, followed by the aquic (2.45%) and xeric (2.10%) classes, respectively. Soils in dense forestlands had the highest SOC contents, whereas soils of younger geological age and alluvial fans had lower SOC. The proposed DNN (hidden layers = 7, and size = 50) is a promising algorithm for handling large numbers of auxiliary data at a province-scale, and due to its flexible structure and the ability to extract more information from the auxiliary data surrounding the sampled observations, it had high accuracy for the prediction of the SOC base-line map and minimal uncertainty.

Data synthesis for crop variety evaluation. A review

Crop varieties should fulfill multiple requirements, including agronomic performance and product quality. Variety evaluations depend on data generated from field trials and sensory analyses, performed with different levels of participation from farmers and consumers. Such multi-faceted variety evaluation is expensive and time-consuming; hence, any use of these data should be optimized. Data synthesis can help to take advantage of existing and new data, combining data from different sources and combining it with expert knowledge to produce new information and understanding that supports decision-making. Data synthesis for crop variety evaluation can partly build on extant experiences and methods, but it also requires methodological innovation. We review the elements required to achieve data synthesis for crop variety evaluation, including (1) data types required for crop variety evaluation, (2) main challenges in data management and integration, (3) main global initiatives aiming to solve those challenges, (4) current statistical approaches to combine data for crop variety evaluation and (5) existing data synthesis methods used in evaluation of varieties to combine different datasets from multiple data sources. We conclude that currently available methods have the potential to overcome existing barriers to data synthesis and could set in motion a virtuous cycle that will encourage researchers to share data and collaborate on data-driven research.

Putting deep learning in perspective for pest management scientists

'Deep learning' is causing rapid technological changes in many fields of science, and conjectures about its potential for transforming everyone's work and lives is a matter of great debate. Unfortunately, it is all too easy to apply it as a 'black box' tool with little consideration of its potential limitations, especially when the data it is being applied to is less than perfect. In this Perspective, I try to put deep learning into a broader mechanistic and historical context by showing how it relates to older forms of artificial intelligence; by providing a general explanation of how it operates; and by exploring some of the challenges involved in its implementation. Examples wherein it has been applied to pest management problems are provided to illustrate how the technology works and the challenges deep learning faces. At least in the near term, its biggest impact on agrochemical development seems likely to come in automating the tedious work involved in assessing agrochemical efficacy, but getting there will require major investments in building large, well-curated data sets to work from and in providing the expertise required to assess the resulting model predictions in real-world scenarios. Deep learning may also come to complement the machine learning methodologies already available for use in pesticide discovery and development, but it seems unlikely to supplant them. © 2020 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Stress Distribution Analysis on Hyperspectral Corn Leaf Images for Improved Phenotyping Quality

High-throughput imaging technologies have been developing rapidly for agricultural plant phenotyping purposes. With most of the current crop plant image processing algorithms, the plant canopy pixels are segmented from the images, and the averaged spectrum across the whole canopy is calculated in order to predict the plant’s physiological features. However, the nutrients and stress levels vary significantly across the canopy. For example, it is common to have several times of difference among Soil Plant Analysis Development (SPAD) chlorophyll meter readings of chlorophyll content at different positions on the same leaf. The current plant image processing algorithms cannot provide satisfactory plant measurement quality, as the averaged color cannot characterize the different leaf parts. Meanwhile, the nutrients and stress distribution patterns contain unique features which might provide valuable signals for phenotyping. There is great potential to develop a finer level of image processing algorithm which analyzes the nutrients and stress distributions across the leaf for improved quality of phenotyping measurements. In this paper, a new leaf image processing algorithm based on Random Forest and leaf region rescaling was developed in order to analyze the distribution patterns on the corn leaf. The normalized difference vegetation index (NDVI) was used as an example to demonstrate the improvements of the new algorithm in differentiating between different nitrogen stress levels. With the Random Forest method integrated into the algorithm, the distribution patterns along the corn leaf’s mid-rib direction were successfully modeled and utilized for improved phenotyping quality. The algorithm was tested in a field corn plant phenotyping assay with different genotypes and nitrogen treatments. Compared with the traditional image processing algorithms which average the NDVI (for example) throughout the whole leaf, the new algorithm more clearly differentiates the leaves from different nitrogen treatments and genotypes. We expect that, besides NDVI, the new distribution analysis algorithm could improve the quality of other plant feature measurements in similar ways.

Detecting and Classifying Pests in Crops Using Proximal Images and Machine Learning: A Review

Pest management is among the most important activities in a farm. Monitoring all different species visually may not be effective, especially in large properties. Accordingly, considerable research effort has been spent towards the development of effective ways to remotely monitor potential infestations. A growing number of solutions combine proximal digital images with machine learning techniques, but since species and conditions associated to each study vary considerably, it is difficult to draw a realistic picture of the actual state of the art on the subject. In this context, the objectives of this article are (1) to briefly describe some of the most relevant investigations on the subject of automatic pest detection using proximal digital images and machine learning; (2) to provide a unified overview of the research carried out so far, with special emphasis to research gaps that still linger; (3) to propose some possible targets for future research.

Non-Destructive Early Detection and Quantitative Severity Stage Classification of Tomato Chlorosis Virus (ToCV) Infection in Young Tomato Plants Using Vis–NIR Spectroscopy

Tomato chlorosis virus (ToCV) is a serious, emerging tomato pathogen that has a significant impact on the quality and quantity of tomato production worldwide. Detecting ToCV via means of spectral measurements in an early pre-symptomatic stage offers an alternative to the existing laboratory methods, leading to better disease management in the field. In this study, leaf spectra from healthy and diseased leaves were measured with a spectrometer. The diseased leaves were subjected to RT-qPCR for the detection and quantification of the titer of ToCV. Neighborhood component analysis (NCA) algorithm was employed for the feature selection of the effective wavelengths and the most important vegetation indices out of the 24 that were tested. Two machine learning methods, namely XY-fusion network (XY-F) and multilayer perceptron with automated relevance determination (MLP–ARD), were employed for the estimation of the disease existence and viral load in the tomato leaves. The results showed that before outlier elimination, the MLP–ARD classifier generally outperformed the XY-F network with an overall accuracy of 92.1% against 88.3% for the XY-F. Outlier elimination contributed to the performance of the classifiers as the overall accuracy for both XY-F and MLP–ARD reached 100%.

Dataset of annotated food crops and weed images for robotic computer vision control

Weed management technologies that can identify weeds and distinguish them from crops are in need of artificial intelligence solutions based on a computer vision approach, to enable the development of precisely targeted and autonomous robotic weed management systems. A prerequisite of such systems is to create robust and reliable object detection that can unambiguously distinguish weed from food crops. One of the essential steps towards precision agriculture is using annotated images to train convolutional neural networks to distinguish weed from food crops, which can be later followed using mechanical weed removal or selected spraying of herbicides. In this data paper, we propose an open-access dataset with manually annotated images for weed detection. The dataset is composed of 1118 images in which 6 food crops and 8 weed species are identified, altogether 7853 annotations were made in total. Three RGB digital cameras were used for image capturing: Intel RealSense D435, Canon EOS 800D, and Sony W800. The images were taken on food crops and weeds grown in controlled environment and field conditions at different growth stages

Bayesian Sigmoid-Type Time Series Forecasting with Missing Data for Greenhouse Crops

This paper follows an integrated approach of Internet of Things based sensing and machine learning for crop growth prediction in agriculture. A Dynamic Bayesian Network (DBN) relates crop growth associated measurement data to environmental control data via hidden states. The measurement data, having (non-linear) sigmoid-type dynamics, are instances of the two classes observed and missing, respectively. Considering that the time series of the logistic sigmoid function is the solution to a reciprocal linear dynamic model, the exact expectation-maximization algorithm can be applied to infer the hidden states and to learn the parameters of the model. At iterative convergence, the parameter estimates are then used to derive a predictor of the measurement data several days ahead. To evaluate the performance of the proposed DBN, we followed three cultivation cycles of micro-tomatoes (MicroTom) in a mini-greenhouse. The environmental parameters were temperature, converted into Growing Degree Days (GDD), and the solar irradiance, both at a daily granularity. The measurement data were Leaf Area Index (LAI) and Evapotranspiration (ET). Although measurement data were only available scarcely, it turned out that high quality measurement data predictions were possible up to three weeks ahead.

Relative Radiometric Calibration Using Tie Points and Optimal Path Selection for UAV Images

As the use of unmanned aerial vehicle (UAV) images rapidly increases so does the need for precise radiometric calibration. For UAV images, relative radiometric calibration is required in addition to the traditional vicarious radiometric calibration due to the small field of view. For relative radiometric calibration, some UAVs install irradiance sensors, but most do not. For UAVs without them, an intelligent scheme for relative radiometric calibration must be applied. In this study, a relative radiometric calibration method is proposed to improve the quality of a reflectance map without irradiance measurements. The proposed method, termed relative calibration by the optimal path (RCOP), uses tie points acquired during geometric calibration to define the optimal paths. A calibrated image from RCOP was compared to validation data calibrated with irradiance measurements. As a result, the RCOP method produces seamless mosaicked images with uniform brightness and reflectance patterns. Therefore, the proposed method can be used as a precise relative radiometric calibration method for UAV images.

Systematic Mapping Study on Remote Sensing in Agriculture

The area of remote sensing techniques in agriculture has reached a significant degree of development and maturity, with numerous journals, conferences, and organizations specialized in it. Moreover, many review papers are available in the literature. The present work describes a literature review that adopts the form of a systematic mapping study, following a formal methodology. Eight mapping questions were defined, analyzing the main types of research, techniques, platforms, topics, and spectral information. A predefined search string was applied in the Scopus database, obtaining 1590 candidate papers. Afterwards, the most relevant 106 papers were selected, considering those with more than six citations per year. These are analyzed in more detail, answering the mapping questions for each paper. In this way, the current trends and new opportunities are discovered. As a result, increasing interest in the area has been observed since 2000; the most frequently addressed problems are those related to parameter estimation, growth vigor, and water usage, using classification techniques, that are mostly applied on RGB and hyperspectral images, captured from drones and satellites. A general recommendation that emerges from this study is to build on existing resources, such as agricultural image datasets, public satellite imagery, and deep learning toolkits.

Machine Learning-Based Predictive Farmland Optimization and Crop Monitoring System

E-agriculture is the integration of technology and digital mechanisms into agricultural processes for more efficient output. This study provided a machine learning-aided mobile system for farmland optimization, using various inputs such as location, crop type, soil type, soil pH, and spacing. Random forest algorithm and BigML were employed to analyze and classify datasets containing crop features that generated subclasses based on random crop feature parameters. The subclasses were further grouped into three main classes to match the crops using data from the companion crops. The study concluded that the approach aided decision making and also assisted in the design of a mobile application using Appery.io. This Appery.io then took in some user input parameters, thereby offering various optimization sets. It was also deduced that the system led to users' optimization of information when implemented on their farmlands.

Predicting Biomass and Yield in a Tomato Phenotyping Experiment Using UAV Imagery and Random Forest

Biomass and yield are key variables for assessing the production and performance of agricultural systems. Modeling and predicting the biomass and yield of individual plants at the farm scale represents a major challenge in precision agriculture, particularly when salinity and other abiotic stresses may play a role. Here, we evaluate a diversity panel of the wild tomato species (Solanum pimpinellifolium) through both field and unmanned aerial vehicle (UAV)-based phenotyping of 600 control and 600 salt-treated plants. The study objective was to predict fresh shoot mass, tomato fruit numbers, and yield mass at harvest based on a range of variables derived from the UAV imagery. UAV-based red-green-blue (RGB) imageries collected 1, 2, 4, 6, 7, and 8 weeks before harvest were also used to determine if prediction accuracies varied between control and salt-treated plants. Multispectral UAV-based imagery was also collected 1 and 2 weeks prior to harvest to further explore predictive insights. In order to estimate the end of season biomass and yield, a random forest machine learning approach was implemented using UAV-imagery-derived predictors as input variables. Shape features derived from the UAV, such as plant area, border length, width, and length, were found to have the highest importance in the predictions, followed by vegetation indices and the entropy texture measure. The multispectral UAV imagery collected 2 weeks prior to harvest produced the highest explained variances for fresh shoot mass (87.95%), fruit numbers (63.88%), and yield mass per plant (66.51%). The RGB UAV imagery produced very similar results to those of the multispectral UAV dataset, with the explained variance reducing as a function of increasing time to harvest. The results showed that predicting the yield of salt-stressed plants produced higher accuracies when the models excluded control plants, whereas predicting the yield of control plants was not affected by the inclusion of salt-stressed plants within the models. This research demonstrates that it is possible to predict the average biomass and yield up to 8 weeks prior to harvest within 4.23% of field-based measurements and up to 4 weeks prior to harvest at the individual plant level. Results from this work may be useful in providing guidance for yield forecasting of healthy and salt-stressed tomato plants, which in turn may inform growing practices, logistical planning, and sales operations.

Segmenting Purple Rapeseed Leaves in the Field from UAV RGB Imagery Using Deep Learning as an Auxiliary Means for Nitrogen Stress Detection

Crop leaf purpling is a common phenotypic change when plants are subject to some biotic and abiotic stresses during their growth. The extraction of purple leaves can monitor crop stresses as an apparent trait and meanwhile contributes to crop phenotype analysis, monitoring, and yield estimation. Due to the complexity of the field environment as well as differences in size, shape, texture, and color gradation among the leaves, purple leaf segmentation is difficult. In this study, we used a U-Net model for segmenting purple rapeseed leaves during the seedling stage based on unmanned aerial vehicle (UAV) RGB imagery at the pixel level. With the limited spatial resolution of rapeseed images acquired by UAV and small object size, the input patch size was carefully selected. Experiments showed that the U-Net model with the patch size of 256 × 256 pixels obtained better and more stable results with a F-measure of 90.29% and an Intersection of Union (IoU) of 82.41%. To further explore the influence of image spatial resolution, we evaluated the performance of the U-Net model with different image resolutions and patch sizes. The U-Net model performed better compared with four other commonly used image segmentation approaches comprising support vector machine, random forest, HSeg, and SegNet. Moreover, regression analysis was performed between the purple rapeseed leaf ratios and the measured N content. The negative exponential model had a coefficient of determination (R²) of 0.858, thereby explaining much of the rapeseed leaf purpling in this study. This purple leaf phenotype could be an auxiliary means for monitoring crop growth status so that crops could be managed in a timely and effective manner when nitrogen stress occurs. Results demonstrate that the U-Net model is a robust method for purple rapeseed leaf segmentation and that the accurate segmentation of purple leaves provides a new method for crop nitrogen stress monitoring.

Spatially and Temporally Explicit Life Cycle Environmental Impacts of Soybean Production in the U.S. Midwest

Understanding spatially and temporally explicit life cycle environmental impacts is critical for designing sustainable supply chains for biofuel and animal sectors. However, annual life cycle environmental impacts of crop production at county scale across mutiple years are lacking. To address this knowledge gap, this study used a combination of Environmental Policy Integrated Climate and process-based life cycle assessment models to quantify life cycle global warming (GWP), eutrophication (EU) and acidification (AD) impacts of soybean production in nearly 1000 Midwest counties yr-1 over 9 years. Sequentially, a machine learning approach was applied to identify the top influential factors among soil, climate, and farming practices, which drive the spatial and temporal heterogeneity of life cycle environmental impacts. The results indicated that significant variations existed in life cycle GWP, EU, and AD among counties and across years. Life cycle GWP impacts ranged from -11.4 to 22.0 kg CO2-eq kg soybean-1, whereas life cycle EU and AD impacts varied by factors of 302 and 44, respectively. Nitrogen application rates, temperature in March and soil texture were the top influencing factors for life cycle GWP impacts. In contrast, soil organic content and nitrogen application rate were the top influencing factors for life cycle EU and AD impacts.

Harnessing genomics to fast-track genetic improvement in aquaculture

Aquaculture is the fastest-growing farmed food sector and will soon become the primary source of fish and shellfish for human diets. In contrast to crop and livestock production, aquaculture production is derived from numerous, exceptionally diverse species that are typically in the early stages of domestication. Genetic improvement of production traits via well-designed, managed breeding programmes has great potential to help meet the rising seafood demand driven by human population growth. Supported by continuous advances in sequencing and bioinformatics, genomics is increasingly being applied across the broad range of aquaculture species and at all stages of the domestication process to optimize selective breeding. In the future, combining genomic selection with biotechnological innovations, such as genome editing and surrogate broodstock technologies, may further expedite genetic improvement in aquaculture.

Performances of the LBP Based Algorithm over CNN Models for Detecting Crops and Weeds with Similar Morphologies

Weed invasions pose a threat to agricultural productivity. Weed recognition and detection play an important role in controlling weeds. The challenging problem of weed detection is how to discriminate between crops and weeds with a similar morphology under natural field conditions such as occlusion, varying lighting conditions, and different growth stages. In this paper, we evaluate a novel algorithm, filtered Local Binary Patterns with contour masks and coefficient k (k-FLBPCM), for discriminating between morphologically similar crops and weeds, which shows significant advantages, in both model size and accuracy, over state-of-the-art deep convolutional neural network (CNN) models such as VGG-16, VGG-19, ResNet-50 and InceptionV3. The experimental results on the "bccr-segset" dataset in the laboratory testbed setting show that the accuracy of CNN models with fine-tuned hyper-parameters is slightly higher than the k-FLBPCM method, while the accuracy of the k-FLBPCM algorithm is higher than the CNN models (except for VGG-16) for the more realistic "fieldtrip_can_weeds" dataset collected from real-world agricultural fields. However, the CNN models require a large amount of labelled samples for the training process. We conducted another experiment based on training with crop images at mature stages and testing at early stages. The k-FLBPCM method outperformed the state-of-the-art CNN models in recognizing small leaf shapes at early growth stages, with error rates an order of magnitude lower than CNN models for canola-radish (crop-weed) discrimination using a subset extracted from the "bccr-segset" dataset, and for the "mixed-plants" dataset. Moreover, the real-time weed-plant discrimination time attained with the k-FLBPCM algorithm is approximately 0.223 ms per image for the laboratory dataset and 0.346 ms per image for the field dataset, and this is an order of magnitude faster than that of CNN models.

Forecasting beef production and quality using large-scale integrated data from Brazil

With agriculture rapidly becoming a data-driven field, it is imperative to extract useful information from large data collections to optimize the production systems. We compared the efficacy of regression (linear regression or generalized linear regression [GLR] for continuous or categorical outcomes, respectively), random forests (RF) and multilayer neural networks (NN) to predict beef carcass weight (CW), age when finished (AS), fat deposition (FD), and carcass quality (CQ). The data analyzed contained information on over 4 million beef cattle from 5,204 farms, corresponding to 4.3% of Brazil's national production between 2014 and 2016. Explanatory variables were integrated from different data sources and encompassed animal traits, participation in a technical advising program, nutritional products sold to farms, economic variables related to beef production, month when finished, soil fertility, and climate in the location in which animals were raised. The training set was composed of information collected in 2014 and 2015, while the testing set had information recorded in 2016. After parameter tuning for each algorithm, models were used to predict the testing set. The best model to predict CW and AS was RF (CW: predicted root mean square error = 0.65, R2 = 0.61, and mean absolute error = 0.49; AS: accuracy = 28.7%, Cohen's kappa coefficient [Kappa] = 0.08). While the best approach for FD and CQ was GLR (accuracy = 45.7%, Kappa = 0.05, and accuracy = 58.7%, Kappa = 0.09, respectively). Across all models, there was a tendency for better performance with RF and regression and worse with NN. Animal category, nutritional plan, cattle sales price, participation in a technical advising program, and climate and soil in which animals were raised were deemed important for prediction of meat production and quality with regression and RF. The development of strategies for prediction of livestock production using real-world large-scale data will be core to projecting future trends and optimizing the allocation of resources at all levels of the production chain, rendering animal production more sustainable. Despite beef cattle production being a complex system, this analysis shows that by integrating different sources of data it is possible to forecast meat production and quality at the national level with moderate-high levels of accuracy.