Revolutionizing Agritech with Deep Learning-Enhanced Remote Sensing for Precision Agricultural Management

Many remote sensing applications are devoted to the agricultural sector. Representative case studies are presented in the special issue “Advances in Remote Sensing of Agriculture”. To complement the examples published within the special issue, a few main applications with regional to global focus were selected for this review, where remote sensing contributions are traditionally strong. The selected applications are put in the context of the global challenges the agricultural sector is facing: minimizing the environmental impact, while increasing production and productivity. Five different applications have been selected, which are illustrated and described: (1) biomass and yield estimation, (2) vegetation vigor and drought stress monitoring, (3) assessment of crop phenological development, (4) crop acreage estimation and cropland mapping and (5) mapping of disturbances and land use/land cover (LULC) changes. Many other applications exist, such as precision agriculture and irrigation management (see other special issues of this journal), but were not included to keep the paper concise. The paper starts with an overview of the main agricultural challenges. This section is followed by a brief overview of existing operational monitoring systems. Finally, in the main part of the paper, the mentioned applications are described and illustrated. The review concludes with some key recommendations.

* About the Editor * Contributors * Foreword (M. S. Swaminathan) * Preface * Acknowledgments * PART I: PRINCIPLES, TECHNOLOGIES, AND MANAGEMENT ISSUES * Chapter 1. Precision Agriculture: An Overview (Ancha Srinivasan) * Introduction * Basics of Precision Agriculture * Tools for Implementation of Precision Agriculture * Current Status, Uncertainties, and Future Trends * Epilogue * Chapter 2. The Role of Technology in the Emergence and Current Status of Precision Agriculture (John V. Stafford) * The Beginnings of Precision Agriculture * The Basis for Precision Agriculture: Information Technology * Spatial Location * Basics of GPS * Information Acquisition: Sensors * Crop Condition * Weed Detection * Grain Yield * Grain Quality * Environment * Assembling and Interpreting Information * Utilizing Information: Application and Control * Agrochemicals * Patch Spraying: Philosophy of Approach * Fertilizers * The Role of Precision Agriculture in the Future of Agriculture-Technological Developments * Chapter 3. Soil Sensors for Precision Farming (Sakae Shibusawa) * Introduction * Current Developments and Use of Soil Sensors * Future Development and Prospects * Conclusions * Chapter 4. Site-Specific Nutrient Management: Objectives, Current Status, and Future Research Needs (Silvia Haneklaus and Ewald Schnug) * Introduction * Origins of SSNM * Data Sources for SSNM * Decision Making for SSNM * SSNM for Different Nutrient Sources * Interaction of SSNM with Other PA Measures in the Field * Quality Aspects * Economic, Ecological, and Social Impacts of SSNM * Future Research Needs * Chapter 5. Precision Water Management: Current Realities, Possibilities, and Trends (Carl R. Camp, E. John Sadler, and Robert G. Evans) * Introduction * Current Status * Irrigation Application and System Control * Auxiliary System Components * Management Zones * Applications and Justifications * Current Trends * Cost-Benefit Issues * Future Directions * Conclusions * Chapter 6. Site-Specific Weed Management (Roland Gerhards and Svend Christensen) * Introduction * Weed Distribution in the Field * Stability of Weed Populations * Weed Monitoring * Decision Making * Site-Specific Herbicide Application * Site-Specific Weed Control * Future Directions * Chapter 7. Site-Specific Management of Crop Diseases (Karsten D. Bjerre, Lise N. Jorgensen, and Jorgen E. Olesen) * Introduction * The Disease Management Arena * IPM Strategies for Disease Control * Site-Specific Disease Control: The Next Step in the Evolution of Disease Management * Effects of Diseases and Spatial Variability on Crop Growth * Technology for Site-Specific Disease Management * Perspectives * Chapter 8. Site-Specific Management of Plant-Parasitic Nematodes (Robert A. Dunn, Jimmy R. Rich, and Richard E. Baird) * Introduction * General Nematode Biology * Diagnosing Nematode Problems * Principles of Nematode Management--Nonchemical * Nematicides * Variable-Rate Nematicide Application * Chapter 9. Site-Specific Measurement and Management of Grain Quality (Piet Reyns, Josse De Baerdemaeker, Ludo Vanongeval, and Maarten Geypens) * Introduction * Quality Factors and Their Measurement * On-Line Quality Measurements * Influence of Plant Nutrition on the Quality of Cereal Crops * Grain Quality and Crop Management * Site-Specific Crop Quantity and Quality Management * Conclusions * PART II: APPLICATIONS IN CROPS AND CROPPING SYSTEMS * Chapter 10. Site-Specific Rice Management (Alvaro Roel, G. Stuart Pettygrove, and Richard E. Plant) * Introduction * Quantifying Spatial Variability and Its Causes * Discussion * Chapter 11. Precision Agriculture Management Progress and Prospects for Corn/Soybean Systems in the Midwestern United States (Thomas S. Colvin) * Introduction * Experimentation in Central Iowa * Availability of Yield Monitors and Site-Specific Soil Testing * Other Benefits of Yield Monitors * Status of Soil Sampling * Profitability * Environmental Issues * The Human Side of Precision Agriculture * The Need for Future Research * Chapter 12. Site-Specific Management of Cotton Production in the United States (Richard M. Johnson, Judith M. Bradow, and Anne F. Wrona) * Introduction * Soil Informational Layer * Crop Informational Layer * Remote Sensing Informational Layer * Integration of Informational Layers * Acceptance of Site-Specific Management by Cotton Producers * Chapter 13. Potential of Precision Farming with Potatoes (Colin McKenzie and Shelley A. Woods) * Introduction * Nutrient Management * Remote Sensing * Nematodes * Insects * Weed Control * Harvesting and Seeding Equipment * Soil Salinity * Field Scale Experimentation * Problems Hindering the Adoption of Precision Farming by the Potato Industry * Conclusions * Chapter 14. Site-Specific Management in Sugarbeet (David W. Franzen) * Properties of Sugarbeet Favorable to Site-Specific Nutrient Management * Zone Management of Nutrients * Profitability of Using Site-Specific Nitrogen Management in Sugarbeet * Use of Imagery from Sugarbeet to Modify Nitrogen Recommendations to Subsequent Crops * Conclusions * Chapter 15. Application of Remote Sensing and Ecosystem Modeling in Vineyard Management (Ramakrishna R. Nemani, Lee F. Johnson, and Michael A. White) * Introduction * The Vineyard As an Ecosystem * Tools in Vineyard Management * Conclusions * Chapter 16. Site-Specific Management from a Cropping System Perspective (David E. Clay, Sharon A. Clay, and Gregg Carlson) * Introduction * Understanding Yield Variability * Managing Yield Variability * Conclusions * PART III: CURRENT STATUS * Chapter 17. Africa (W. T. (Wimpie) Nell, Ntsikane Maine, and P. M. Basson) * Introduction * Climatic Conditions * Background of Agriculture * Site-Specific Management * Precision Agriculture * Constraints in the Adoption of Precision Agriculture and Site-Specific Management Technologies * Research on Precision Agriculture in South Africa * Prospects for Precision Agriculture * Chapter 18. Asia (Ancha Srinivasan) * Introduction * Spatial Variability in Asian Farms * Drivers and Opportunities for Adoption of Precision Farming * Current Status in Selected Countries * Constraints and Approaches for Adoption * Implications for Adoption in Asia * Future Action * Conclusions * Chapter 19. Australia (Simon E. Cook, Matthew L. Adams, Robert G. V. Bramley, and Brett M. Whelan) * Introduction * What Precision Agriculture Means in Australia * Demand for Precision Agriculture in Australia: The Battle for Sustainability Needs Accurate and Relevant Information * Methods Used in Australia * Applications in the Grains, Cotton, Wine, and Sugar Industries * Impediments to Adoption * Conclusions * Chapter 20. Europe (Simon Blackmore, Hans W. Greipentrog, Soren M. Pedersen, and Spyros Fountas) * Introduction * The Current Situation in European Farming * Precision Farming Research in Europe * Variability and Management * Technology-Led Opportunities * Issues of Adoption and Farmer Attitudes * Future Research * Conclusions * Chapter 21. Argentina (Rodolfo Bongiovanni and Jess Lowenberg-DeBoer) * Introduction * Argentine Agriculture * Current Status * Factors That Favor Adoption * Factors That Discourage Adoption * Prospects * Challenges * Chapter 22. Brazil (Glaucio Roloff and Daniele Focht) * Introduction * A Brief History of Precision Agriculture in Brazil * Precision Agriculture on Highly Weathered Soils * Managing Variability * Precision Agriculture for Specific Crops * Conclusions * Index * Reference Notes Included

In the fields of agriculture and forestry, the Normalized Difference Vegetation Index (NDVI) is a critical indicator for assessing the physiological state of plants. Traditional imaging sensors can only collect two-dimensional vegetation distribution data, while dual-wavelength LiDAR technology offers the capability to capture vertical distribution information, which is essential for forest structure recovery and precision agriculture management. However, existing LiDAR systems face challenges in detecting echoes at two wavelengths, typically relying on multiple detectors or array sensors, leading to high costs, bulky systems, and slow detection rates. This study introduces a time-stretched method to separate two laser wavelengths in the time dimension, enabling a more cost-effective and efficient dual-spectral (600 nm and 800 nm) LiDAR system. Utilizing a supercontinuum laser and a single-pixel detector, the system incorporates specifically designed time-stretched transmission optics, enhancing the efficiency of NDVI data collection. We validated the ranging performance of the system, achieving an accuracy of approximately 3 mm by collecting data with a high sampling rate oscilloscope. Furthermore, by detecting branches, soil, and leaves in various health conditions, we evaluated the system’s performance. The dual-wavelength LiDAR can detect variations in NDVI due to differences in chlorophyll concentration and water content. Additionally, we used the radar equation to analyze the actual scene, clarifying the impact of the incidence angle on reflectance and NDVI. Scanning the Red Sumach, we obtained its NDVI distribution, demonstrating its physical characteristics. In conclusion, the proposed dual-wavelength LiDAR based on the time-stretched method has proven effective in agricultural and forestry applications, offering a new technological approach for future precision agriculture and forest management.

The agricultural sector necessitates automated identification and analysis of rice diseases to conserve financial and other resources, mitigate yield loss, enhance processing efficiency, and secure healthy crop harvests. Rice is an essential commodity for global food security; however, it is very vulnerable to numerous illnesses that can markedly diminish productivity. Timely identification and precise forecasting of these diseases are crucial for reducing losses. Conventional image-based disease detection techniques frequently utilise Convolutional Neural Networks (CNNs) to extract spatial information; however, they inadequately account for the temporal evolution of diseases, which is essential for efficient monitoring and diagnosis. This research proposes a hybrid model that integrates a Hierarchical Convolutional Recurrent Neural Network (HCRNN) with Long Short-Term Memory (LSTM) networks for the prediction of rice plant illnesses. The HCRNN extracts multi-scale spatial characteristics from rice plant pictures, whilst the LSTM network models temporal relationships in disease progression, hence augmenting predicting capabilities. This integrated methodology enhances performance by integrating spatial and temporal information. We assessed the model using a dataset of rice plant leaf pictures impacted by multiple diseases, including bacterial leaf blight, blast, and sheath blight. The proposed model exhibited enhanced performance, with an accuracy of 98.5%, above that of conventional CNN-based models. This method also resolves the challenge of limited datasets by accurately tracking disease development across time. The findings indicate that the integration of HCRNN with LSTM establishes a resilient framework for predicting rice diseases. The suggested approach is adaptable to additional crops and disease categories, providing a scalable solution for precision agriculture and disease management. Subsequent efforts will concentrate on incorporating environmental variables, including soil and meteorological data, to augment predictive accuracy.

This study introduces an innovative approach to predictive crop management in precision agriculture by integrating deep learning with multispectral remote sensing technologies. The research aims to develop a framework that combines multispectral data from field sensors, UAVs, and satellites with a deep learning model based on a multimodal architecture incorporating adaptive transfer learning and attention mechanisms. Data were collected over two growing seasons and underwent preprocessing, vegetation feature extraction, and model training and validation. The proposed deep learning model significantly outperformed traditional machine learning algorithms such as Random Forest and Support Vector Machines, achieving up to 97.8% accuracy in crop classification. Predicted crop conditions and yield estimates showed a strong correlation with actual field data (r = 0.89; RMSE = 0.12). Field implementation of the predictive system indicated potential increases in crop yield by 18% and reductions in agricultural input usage by 28%. These results highlight the potential of deep learning and multispectral data integration to enhance decision-making, resource efficiency, and sustainability in precision farming. Furthermore, the approach demonstrates strong scalability for different crop types and geographical regions, providing a solid foundation for the digital transformation of agriculture toward a more adaptive and sustainable food production system.

This study addresses the limitations of current non-destructive techniques for assessing tomato quality, such as their high cost, strong dependence on spectroscopic instruments, and difficulty in dynamic monitoring. The study proposes an integrated tomato quality prediction model that combines a Long Short-Term Memory (LSTM)-based environmental predictor, a Gated Recurrent Unit with attention mechanism (GRU-AT) for dynamic maturity prediction, and a Deep Neural Network (DNN)-based quality evaluation module. The LSTM model demonstrated high accuracy in environmental prediction (R2 > 0.9559). The GRU-AT model excelled in color ratio prediction (R2 > 0.86), and the DNN model achieved R2 values exceeding 0.811 for lycopene (LYC), firmness (FI), and soluble solids content (SSC). Experimental results demonstrate that this approach can accurately predict multiple quality parameters using only standard RGB images. In summary, this study provides a low-cost, low-complexity solution enabling real-time, non-destructive monitoring of greenhouse tomato quality, offering a viable pathway for crop quality management in precision agriculture.

In-depth knowledge about soil moisture dynamics is crucial for irrigation management in precision agriculture. This study evaluates the feasibility of high spatial resolution near-infrared remote sensing with unmanned aerial systems for soil moisture estimation to provide decision support for precision irrigation management. A new trapezoid model based on near-infrared transformed reflectance (NTR) and the normalized difference vegetation index (NDVI) is introduced and used for estimation and mapping of root zone soil moisture and plant extractable water. The performance of the proposed approach was evaluated via comparison with ground soil moisture measurements with advanced time domain reflectometry sensors. We found the estimates based on the NTR−NDVI trapezoid model to be highly correlated with the ground soil moisture measurements. We believe that the presented approach shows great potential for farm-scale precision irrigation management but acknowledge that more research for different cropping systems, soil textures, and climatic conditions is needed to make the presented approach viable for the application by crop producers.

Rice is a primary staple food for the world population and there is a strong need to map its cultivation area and monitor its crop status on regional scales. This study was conducted in the Qixing Farm County of the Sanjiang Plain, Northeast China. First, the rice cultivation areas were identified by integrating the remote sensing (RS) classification maps from three dates and the Geographic Information System (GIS) data obtained from a local agency. Specifically, three FORMOSAT-2 (FS-2) images captured during the growing season in 2009 and a GIS topographic map were combined using a knowledge-based classification method. A highly accurate classification map (overall accuracy = 91.6%) was generated based on this Multi-Data-Approach (MDA). Secondly, measured agronomic variables that include biomass, leaf area index (LAI), plant nitrogen (N) concentration and plant N uptake were correlated with the date-specific FS-2 image spectra using stepwise multiple linear regression models. The best model validation results with a relative error (RE) of 8.9% were found in the biomass regression model at the phenological stage of heading. The best index of agreement (IA) value of 0.85 with an RE of 13.6% was found in the LAI model, also at the heading stage. For plant N uptake estimation, the most accurate model was again achieved at the heading stage with an RE of 11% and an IA value of 0.77; however, for plant N concentration estimation, the model performance was best at the booting stage. Finally, the regression models were applied to the identified rice areas to map the within-field variability of the four agronomic variables at different growth stages for the Qixing Farm County. The results provide detailed spatial information on the within-field variability on a regional scale, which is critical for effective field management in precision agriculture.

This study explores the role of precision agriculture, climate-smart farming, and sustainable supply chain management in boosting agricultural productivity in 2024. The primary objective is to qualitatively analyze the literature to understand how these innovative practices contribute to enhancing agricultural productivity and sustainability. The research employs a qualitative literature review methodology, synthesizing findings from academic articles, industry reports, case studies, and empirical studies to provide a comprehensive overview of the current state of knowledge in this field. The literature review methodology involves systematically collecting and analyzing scholarly sources that discuss various aspects of precision agriculture, climate-smart farming, and sustainable supply chain management. The study categorizes the literature into key themes, such as the technological advancements in precision agriculture, the principles and practices of climate-smart farming, and the impact of sustainable supply chain management on agricultural productivity and sustainability. Thematic analysis is used to identify patterns and trends in how these practices interact to influence agricultural outcomes. The findings indicate that precision agriculture, through the use of advanced technologies like GPS, IoT, and AI, enables farmers to optimize field-level management regarding crop farming. This leads to increased yield, reduced waste, and better resource utilization. Climate-smart farming practices, including crop diversification, improved irrigation techniques, and soil health management, are essential for adapting to climate change and mitigating its impacts on agriculture. Sustainable supply chain management ensures that agricultural products are produced, processed, and distributed in ways that minimize environmental impact and enhance economic viability.

Weed and crop species classification using computer vision and deep learning technologies in greenhouse conditions

Real-time plant disease diagnosis employs new technologies to identify and detect plant diseases while they occur, thus allowing a rapid response that reduces crop loss and improves healthier agricultural practices. This work improves plant health monitoring using early detection to maximize yield and reduce loss. Typical procedures for diagnosing plant diseases involve sampling or visual inspection and are slow, labor-intensive, and subject to human error. These procedures are not suitable for widespread adoption in the field of crop systems where the scale of diagnostics requires real-time, scalable, and accurate reporting of problems. The Plant Disease Diagnosis using Deep Learning (PDD-DL) framework, through Convolutional Neural Networks (CNNs), analyzes plant images to automatically diagnose plant diseases in real time. This model is faster, more trustworthy, and more scalable in diagnosis than traditional methods of diagnosis. The research presents the validation of the model based on common, popular crops; however, the application includes a wide array of crops. The system may be retrained for specific disease classes depending on agricultural requirements. CNNs will certainly provide effective image analysis, accurately differentiating healthy from sick plants, and permitting continuous monitoring for preemptive measures in the classification of plant diseases. The proposed model performed with an overall accuracy of 98.32%, precision score of 97.85%, recall value of 98.14%, F1-score of 97.99%, and real-time inference speed of 42.6 ms per image. As a result, the study’s findings improve accuracy and speed in diagnosing plant disease, which aids in precision agriculture and sustainable plant health management.

In Pedology and agricultural management, soil moisture is crucial for maintaining the physicochemical, biological, agronomical, ecological, hydrographical, and geomorphic soil features. The framework for managing irrigation and making efficient use of available water resources is provided by the soil moisture detection system. It contributes significantly to Precision Agriculture (PA) by constant monitoring of humidity and moisture content data in real-time. The high cost, the necessity of site-specific measurement, poor performance, and small sampling capacity of soil humidity sensors limit their applications. The objective is to investigate the effectiveness of all soil moisture monitoring systems in addition to the developments in novel detection methods and to assess their applicability in agricultural soil management. Based on their performance and design, a study of the benefits and drawbacks of soil moisture detectors is conducted. The development of sensor systems has led to an improvement in detection approaches by utilizing a set of technologies, including Wireless Sensor Networks (WSN), Internet of Things (IoT) and Remote Sensing (RS). The diverse RS, IoT, and WSN techniques utilized in Precision Agriculture are covered in this overview, along with their effects on the progress of "smart agriculture." This paper conducts a rigorous review of the WSN, RS, and agricultural IoT research status. To achieve smart and intelligent agricultural production, the study also focuses on the optimization of environmental parameters, such as soil property monitoring and irrigation management. Moreover, the issues and difficulties associated with detecting soil moisture are examined, and a projection for the future growth of agricultural IoT, RS, and WSNs is provided. Finally, this review discusses how novel technologies are potentially applied to detect soil moisture.

In this paper, the primary focus is of Slot Tagging of Gujarat Dialogue, which enables the Gujarati language communication between human and machine, allowing machines to perform given task and provide desired output. The accuracy of tagging entirely depends on bifurcation of slots and word embedding. It is also very challenging for a researcher to do proper slot tagging as dialogue and speech differs from human to human, which makes the slot tagging methodology more complex. Various deep learning models are available for slot tagging for the researchers, however, in the instant paper it mainly focuses on Long Short-Term Memory (LSTM), Convolutional Neural Network - Long Short-Term Memory (CNN-LSTM) and Long Short-Term Memory – Conditional Random Field (LSTM-CRF), Bidirectional Long Short-Term Memory (BiLSTM), Convolutional Neural Network - Bidirectional Long Short-Term Memory (CNN-BiLSTM) and Bidirectional Long Short-Term Memory – Conditional Random Field (BiLSTM-CRF). While comparing the above models with each other, it is observed that BiLSTM models performs better than LSTM models by a variation ~2% of its F1-measure, as it contains an additional layer which formulates the word string to traverse from backward to forward. Within BiLSTM models, BiLSTM-CRF has outperformed other two Bi-LSTM models. Its F1-measure is better than CNN-BiLSTM by 1.2% and BiLSTM by 2.4%.KeywordsSpoken Language Understanding (SLU)Long Short-Term Memory (LSTM)Slot taggingBidirectional Long Short-Term Memory (BiLSTM)Convolutional Neural Network - Bidirectional Long Short-Term Memory (CNN-BiLSTM)Bidirectional Long Short-Term Memory (BiLSTM-CRF)

Image has become one of the important carriers of visual information because of its large amount of information, easy to spread and store, and strong sense of sense. At the same time, the quality of image is also related to the completeness and accuracy of information transmission. This research mainly discusses the superresolution reconstruction of remote sensing images based on the middle layer supervised convolutional neural network. This paper designs a convolutional neural network with middle layer supervision. There are 16 layers in total, and the seventh layer is designed as an intermediate supervision layer. At present, there are many researches on traditional superresolution reconstruction algorithms and convolutional neural networks, but there are few researches that combine the two together. Convolutional neural network can obtain the high-frequency features of the image and strengthen the detailed information; so, it is necessary to study its application in image reconstruction. This article will separately describe the current research status of image superresolution reconstruction and convolutional neural networks. The middle supervision layer defines the error function of the supervision layer, which is used to optimize the error back propagation mechanism of the convolutional neural network to improve the disappearance of the gradient of the deep convolutional neural network. The algorithm training is mainly divided into four stages: the original remote sensing image preprocessing, the remote sensing image temporal feature extraction stage, the remote sensing image spatial feature extraction stage, and the remote sensing image reconstruction output layer. The last layer of the network draws on the single-frame remote sensing image SRCNN algorithm. The output layer overlaps and adds the remote sensing images of the previous layer, averages the overlapped blocks, eliminates the block effect, and finally obtains high-resolution remote sensing images, which is also equivalent to filter operation. In order to allow users to compare the superresolution effect of remote sensing images more clearly, this paper uses the Qt5 interface library to implement the user interface of the remote sensing image superresolution software platform and uses the intermediate layer convolutional neural network and the remote sensing image superresolution reconstruction algorithm proposed in this paper. When the training epoch reaches 35 times, the network has converged. At this time, the loss function converges to 0.017, and the cumulative time is about 8 hours. This research helps to improve the visual effects of remote sensing images.

Recent advances in optical sensor technologies and Geoinformatics, can support very large scale high definition, used for multispectral and panchromatic images. This capability allows the use of remote sensing for the observation of complex earth ecosystems. Application areas include, sustainability of biodiversity, precision agriculture, land, crops and parasites management. Moreover, it supports advanced quantitative studies of biophysical and biogeochemical cycles, in costal or inland waters. The requirement for precise and effective scene classification, can significantly contribute towards the development of new types of decision support systems. This offers considerable advantages to business, science and engineering. This research paper proposes a novel and effective approach based on geographic object-based scene classification in remote sensing images. More specifically, it introduces an important upgrade of the well-known Residual Neural Network (ResNet) architecture. The omission of some layers in the early stages of training, achieves an effective simplification of the network, by eliminating the “Vanishing Gradient Problem” (VGP) which causes efficiency limitations in other “Deep Learning” (DEL) architectures. The use of the Softmax activation function instead of the Sigmoid in the last layer, is the most important innovation of the proposed system. The ResNet has been trained using the novel AdaBound algorithm that employs dynamic bounds on the employed learning rates. The result is the employment of a smooth transition of the stochastic gradient descent, tackling the noise dispersed points of misclassification with great precision. This is something that other spectral classification methods cannot handle. The proposed algorithm was successfully tested, in scene identification from remote sensing images. This confirms that it could be further used in advanced level processes for Large-Scale Geospatial Data Analysis, such as cross-border classification, recognition and monitoring of certain patterns and multi-sensor data fusion.