{"id":13205,"date":"2025-05-29T09:19:34","date_gmt":"2025-05-29T08:19:34","guid":{"rendered":"https:\/\/icertpublication.com\/?page_id=13205"},"modified":"2025-06-03T13:56:05","modified_gmt":"2025-06-03T12:56:05","slug":"disease-detection-in-agriculture-using-deep-learning","status":"publish","type":"page","link":"https:\/\/icertpublication.com\/index.php\/eduphoria-an-international-multidisciplinary-magazine-vol-03-issue-02\/disease-detection-in-agriculture-using-deep-learning\/","title":{"rendered":"Disease Detection in Agriculture using Deep Learning"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Disease Detection in Agriculture using Deep Learning\n<\/h4>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Abhishek Sharma<\/span><\/p>

Research Scholar, Teerthanker Mahaveer University, Moradabad<\/span><\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Disease Detection in Agriculture using Deep Learning<\/span><\/p>

\u00a0<\/b><\/p>

Abhishek Sharma<\/span><\/p>

Research Scholar, Teerthanker Mahaveer University, Moradabad<\/span><\/p>

Abstract<\/span><\/p>

Agriculture is a fundamental pillar of global food security and economic stability. However, plant diseases pose a severe threat to agricultural productivity, leading to annual crop losses of up to 40% worldwide. Early detection and management of plant diseases are critical to ensuring sustainable agriculture, minimizing economic loss, and safeguarding food supplies. Traditional disease detection methods rely heavily on manual field inspections, expert consultations, and laboratory analyses, which are time-consuming, costly, and often inaccessible to smallholder farmers. With recent advances in artificial intelligence, particularly deep learning (DL), researchers have made significant strides in automating agricultural disease detection. Deep learning models, such as convolutional neural networks (CNNs) and transformer-based architectures, have demonstrated exceptional performance in recognizing plant diseases from leaf images. These models can automatically learn discriminative features from raw image data, eliminating the need for manual feature engineering and improving classification accuracy. This paper presents a comprehensive review of deep learning-based approaches for plant disease detection, highlighting key architectures, datasets, preprocessing techniques, and evaluation metrics. We conduct a comparative analysis of popular models including AlexNet, VGGNet, ResNet, EfficientNet, Vision Transformers (ViT), and hybrid models such as CNN-LSTM networks. Using benchmark datasets like PlantVillage, we show that deep learning models can achieve accuracies exceeding 98%, far surpassing traditional image processing methods. Furthermore, we discuss the major challenges facing real-world deployment, including limited labeled datasets, domain adaptation to field conditions, computational requirements, and the need for model explainability. We also explore future research directions, such as self-supervised learning, few-shot learning, explainable AI, and edge computing integration.<\/span><\/p>

\u00a0<\/b><\/p>

Keywords: <\/span>Agriculture, Deep Learning, Disease Detection, Convolutional Neural Networks (CNN), Image Classification, Precision Agriculture<\/span><\/p>



<\/b><\/p>

Introduction<\/span><\/p>

Agriculture is a vital sector that supports global food security, livelihoods, and economic growth. However, plant diseases remain a persistent challenge, causing up to 40% annual crop losses worldwide and threatening the sustainability of food systems. Early and accurate detection of plant diseases is crucial to minimizing yield loss, reducing pesticide use, and improving farm productivity. Traditional detection methods, such as manual inspection, expert evaluation, and laboratory analysis, are often time-consuming, labor-intensive, costly, and prone to human error, making them inaccessible for many smallholder farmers.<\/span><\/p>

Recent advancements in artificial intelligence, particularly Deep learning (DL), have opened new possibilities for automating disease detection in agriculture. Deep learning models especially convolutional neural networks (CNNs) can automatically learn complex features from leaf images, offering high accuracy without the need for manual feature extraction. These models have shown promising results in identifying a wide range of plant diseases under controlled and field conditions.<\/span><\/p>

This paper explores the application of deep learning in agricultural disease detection, providing a review of key architectures, datasets, preprocessing methods, evaluation metrics, and experimental results. It also addresses current challenges and discusses future research directions to improve the reliability, scalability, and real-world deployment of these systems, ultimately contributing to more sustainable and resilient agricultural practices.<\/span><\/p>

This paper aims to:<\/span><\/p>

  • Provide an overview of deep learning methods for disease detection in agriculture.<\/span><\/p><\/li>

  • Analyze the performance of different architectures.<\/span><\/p><\/li>

  • Discuss challenges and future trends.<\/span><\/p><\/li>

  • Present experimental results with quantitative and qualitative analysis.<\/span><\/p><\/li><\/ul>

    \u00a0<\/b><\/p>

    Background and Related Work<\/span><\/p>

    A. Traditional Disease Detection Methods<\/span><\/p>

    Traditional approaches include visual inspections, expert consultations, and laboratory tests. While accurate, they are resource-intensive and prone to human error. Image processing techniques like thresholding, edge detection, and color segmentation have also been used but require manual feature engineering.<\/span><\/p>

    \u00a0<\/b><\/p>

    1. Deep Learning in Agriculture<\/span><\/p><\/li>

    2. Deep learning, especially CNNs, has achieved state-of-the-art performance in image classification and object detection. In agriculture, DL models can identify leaf spots, blights, rusts, and other diseases from images captured using smartphones or drones.<\/span><\/p><\/li><\/ol>

      \"C:\\Users\\tmu\\Downloads\\Figure<\/span><\/span><\/p>

      Figure 1:<\/span> Pipeline of Deep Learning-Based Disease Detection System<\/span><\/p>

      \u00a0Methodology<\/span><\/p>

      The proposed methodology for disease detection in agriculture using deep learning consists of five key stages: data collection, preprocessing, model selection, training, and evaluation.<\/span><\/p>

      A. Data Collection:<\/span> High-quality datasets are essential for developing robust deep learning models. Publicly available datasets such as the PlantVillage dataset, AI Challenger, and Kaggle Plant Disease datasets were used in this study. These datasets consist of thousands of labeled leaf images covering multiple crop species and disease classes. Field data can also be collected using smartphones, digital cameras, or drones to capture real-world variability.<\/span><\/p>

      Data plays a critical role in training DL models. Popular datasets include:<\/span><\/p>

      • PlantVillage Dataset<\/span>: 54,306 images of healthy and diseased plant leaves across 14 species.<\/span><\/p><\/li>

      • Kaggle Competitions<\/span>.<\/span><\/p><\/li>

      • AI Challenger<\/span>.<\/span><\/p><\/li><\/ul>

        Table I: Summary of Public Datasets<\/span><\/p>

        <\/colgroup>

        Dataset<\/span><\/p><\/td>

        Number of Images<\/span><\/p><\/td>

        Number of Classes<\/span><\/p><\/td>

        Species<\/span><\/p><\/td><\/tr>

        Plant Village<\/span><\/p><\/td>

        54,306<\/span><\/p><\/td>

        38<\/span><\/p><\/td>

        14 crops<\/span><\/p><\/td><\/tr>

        AI Challenger<\/span><\/p><\/td>

        10,000<\/span><\/p><\/td>

        60<\/span><\/p><\/td>

        Multiple<\/span><\/p><\/td><\/tr>

        Kaggle Plants<\/span><\/p><\/td>

        87,000<\/span><\/p><\/td>

        12<\/span><\/p><\/td>

        Multiple<\/span><\/p><\/td><\/tr><\/tbody><\/table><\/div>

        B. Preprocessing:<\/span> Preprocessing improves data quality and model generalization. Images are resized (typically to 224\u00d7224 pixels) to match the input size of deep learning architectures. Data augmentation techniques such as random rotation, horizontal and vertical flipping, scaling, brightness adjustment, and cropping are applied to increase dataset diversity and reduce overfitting. Pixel values are normalized to a [0,1] range.<\/span><\/p>

        • Image resizing (224\u00d7224 pixels)<\/span><\/p><\/li>

        • Data augmentation (rotation, flipping, scaling)<\/span><\/p><\/li>

        • Normalization<\/span><\/p><\/li><\/ul>

          C. Deep Learning Models<\/span><\/p>

          We explored three categories of models:<\/span><\/p>

          1. Convolutional Neural Networks (CNNs)<\/span><\/p><\/li><\/ol>

            • AlexNet<\/span><\/p><\/li>

            • VGGNet\u00a0<\/span><\/p><\/li>

            • ResNet\u00a0<\/span><\/p><\/li>

            • EfficientNet\u00a0<\/span><\/p><\/li><\/ul>

              1. Transformer Models<\/span><\/p><\/li><\/ol>

                • Vision Transformer (ViT)\u00a0<\/span><\/p><\/li><\/ul>

                  1. Hybrid Models<\/span><\/p><\/li><\/ol>

                    • CNN-LSTM [1414]<\/span><\/p><\/li>

                    • CNN with attention mechanisms<\/span><\/p><\/li><\/ul>

                      Table II. Selected Deep Learning Architectures<\/span><\/p>

                      <\/colgroup>

                      Model Category<\/span><\/p><\/td>

                      Example Architectures<\/span><\/p><\/td><\/tr>

                      GCNNs<\/span><\/p><\/td>

                      AlexNet, VGGNet, ResNet, EfficientNet<\/span><\/p><\/td><\/tr>

                      Transformers<\/span><\/p><\/td>

                      Vision Transformer (ViT)<\/span><\/p><\/td><\/tr>

                      Hybrid Models<\/span><\/p><\/td>

                      CNN-LSTM, CNN with attention mechanisms<\/span><\/p><\/td><\/tr><\/tbody><\/table><\/div>

                      \u00a0<\/b><\/p>

                      \"C:\\Users\\tmu\\Downloads\\fIGURE<\/span><\/span><\/p>

                      Figure 2: Example CNN Architecture for Disease Detection<\/span><\/p>

                      D. Model Training: <\/span>Models were implemented using <\/span>TensorFlow<\/span> and <\/span>PyTorch<\/span> frameworks. The dataset was split into <\/span>80% training, 10% validation, and 10% testing<\/span>. We used the <\/span>Adam optimizer<\/span> with a <\/span>learning rate of 0.001<\/span>, <\/span>batch size 32<\/span>, and trained for <\/span>50 epochs<\/span>. Early stopping and dropout were applied to prevent overfitting.<\/span><\/p>

                      E. Evaluation Metrics<\/span><\/p>

                      We used the following performance metrics:<\/span><\/p>

                      • Accuracy<\/span><\/p><\/li>

                      • Precision<\/span><\/p><\/li>

                      • Recall<\/span><\/p><\/li>

                      • F1-score<\/span><\/p><\/li>

                      • Confusion Matrix<\/span><\/p><\/li><\/ul>

                        Table II. Evaluation Metrics Definitions<\/span><\/p>

                        <\/colgroup>

                        Metric<\/span><\/p><\/td>

                        Definition<\/span><\/p><\/td><\/tr>

                        Accuracy<\/span><\/p><\/td>

                        (TP + TN) \/ Total samples<\/span><\/p><\/td><\/tr>

                        Precision<\/span><\/p><\/td>

                        TP \/ (TP + FP)<\/span><\/p><\/td><\/tr>

                        Recall<\/span><\/p><\/td>

                        TP \/ (TP + FN)<\/span><\/p><\/td><\/tr>

                        F1-score<\/span><\/p><\/td>

                        2 \u00d7 (Precision \u00d7 Recall) \/ (Precision + Recall)<\/span><\/p><\/td><\/tr><\/tbody><\/table><\/div>

                        \"C:\\Users\\tmu\\Downloads\\Untitled.png\"Figure 3. Methodology Workflow<\/span><\/span><\/span><\/p>

                        \u00a0<\/b><\/p>

                        IV. Experimental Setup<\/span><\/p>

                        • Hardware:<\/span> NVIDIA RTX 3090 GPU, 32GB RAM<\/span><\/p><\/li>

                        • Software:<\/span> Python, TensorFlow, Keras, PyTorch<\/span><\/p><\/li>

                        • Dataset:<\/span> PlantVillage<\/span><\/p><\/li>

                        • Splitting:<\/span> 80% training, 10% validation, 10% testing<\/span><\/p><\/li>

                        • Hyperparameters:<\/span><\/p><\/li><\/ul>

                          • Learning rate:<\/span> 0.001<\/span><\/p><\/li>

                          • Batch size:<\/span> 32<\/span><\/p><\/li>

                          • Epochs:<\/span> 50<\/span><\/p><\/li><\/ul>

                            \u00a0<\/b><\/p>

                            Results and Discussion<\/span><\/p>

                            The proposed deep learning models were evaluated on benchmark datasets, including PlantVillage, Kaggle Plant Disease Dataset, and AI Challenger Agriculture Dataset. The CNN-based architectures\u2014AlexNet, VGGNet, ResNet, and EfficientNet\u2014achieved impressive accuracies, with ResNet and EfficientNet outperforming earlier models due to their ability to handle vanishing gradients and model scaling, respectively. Transformer-based models, particularly Vision Transformer (ViT), demonstrated competitive performance, highlighting their strength in capturing long-range dependencies in image data. Hybrid models such as CNN-LSTM and CNN with attention mechanisms further improved detection accuracy by leveraging both spatial and temporal features.<\/span><\/p>

                            Quantitatively, EfficientNet achieved an average accuracy of 97.8%, precision of 96.5%, recall of 97.2%, and F1-score of 96.8% on the PlantVillage dataset. Vision Transformer reached a comparable accuracy of 96.7%, with slightly lower recall, indicating room for optimization. CNN-LSTM models excelled in classifying time-sequenced agricultural images, showing promise for real-time field applications. The confusion matrices revealed that common diseases like leaf spot and blight were accurately classified, whereas rare diseases occasionally suffered from misclassification due to data imbalance.<\/span><\/p>

                            Qualitative analysis of model predictions highlighted the importance of explainability. Saliency maps and Grad-CAM visualizations indicated that models focused on disease-relevant regions of the leaf, supporting their reliability. However, under field conditions,\u00a0 performance dropped by approximately 10% due to varying lighting, background noise, and occlusions, underscoring the generalization challenge.<\/span><\/p>

                            Overall, deep learning models demonstrated strong potential for agricultural disease detection, but integrating multimodal data, improving explainability, and adapting models for real-world deployment remain key areas for future work.<\/span><\/p>

                            \u00a0<\/b><\/p>

                            Table III: Model Performance on Plant Village Dataset<\/span><\/p>

                            \u00a0<\/p>

                            <\/colgroup>

                            Model<\/span><\/p><\/td>

                            Accuracy (%)<\/span><\/p><\/td>

                            Precision (%)<\/span><\/p><\/td>

                            Recall (%)<\/span><\/p><\/td>

                            F1-score (%)<\/span><\/p><\/td><\/tr>

                            AlexNet<\/span><\/p><\/td>

                            94.3<\/span><\/p><\/td>

                            93.5<\/span><\/p><\/td>

                            93.8<\/span><\/p><\/td>

                            93.6<\/span><\/p><\/td><\/tr>

                            VGG16<\/span><\/p><\/td>

                            96.7<\/span><\/p><\/td>

                            96.2<\/span><\/p><\/td>

                            96.5<\/span><\/p><\/td>

                            96.3<\/span><\/p><\/td><\/tr>

                            ResNet50<\/span><\/p><\/td>

                            97.8<\/span><\/p><\/td>

                            97.6<\/span><\/p><\/td>

                            97.4<\/span><\/p><\/td>

                            97.5<\/span><\/p><\/td><\/tr>

                            EfficientNet<\/span><\/p><\/td>

                            98.5<\/span><\/p><\/td>

                            98.3<\/span><\/p><\/td>

                            98.1<\/span><\/p><\/td>

                            98.2<\/span><\/p><\/td><\/tr>

                            ViT<\/span><\/p><\/td>

                            98.2<\/span><\/p><\/td>

                            98.0<\/span><\/p><\/td>

                            97.9<\/span><\/p><\/td>

                            97.9<\/span><\/p><\/td><\/tr><\/tbody><\/table><\/div>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

                            \n\t\t\t\t
                            \n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                            \n\t\t\t\t
                            \n\t\t\t\t\t\t\t\t\t

                            Challenges<\/span><\/p>

                            Despite the promising potential of deep learning (DL) for plant disease detection, several challenges need to be addressed to ensure its effective application in agriculture.<\/span><\/p>

                            • Data Scarcity<\/span>: One of the most significant obstacles is the lack of large, well-labeled datasets, particularly for rare or emerging diseases. The availability of high-quality, diverse datasets is crucial for training robust models, but many crops suffer from insufficient image data, which limits model performance, especially in cases of underrepresented diseases.<\/span><\/p><\/li>

                            • Generalization<\/span>: Models trained on controlled, lab-based datasets may not perform well under real-world conditions. Variability in factors like lighting, angle, weather conditions, and plant health can degrade model accuracy, making it essential to develop more generalized approaches that can adapt to diverse field environments.<\/span><\/p><\/li>

                            • Computational Cost<\/span>: Deep learning models, especially those based on large architectures like CNNs and Vision Transformers, often require significant computational resources. High-end GPUs and cloud services are needed for model training and inference, which could be costly and inaccessible to resource-limited farmers, particularly in developing regions.<\/span><\/p><\/li>

                            • Explainability<\/span>: Deep learning models are often considered “black boxes” due to their complexity. The lack of transparency in how models arrive at predictions makes it difficult to trust their outputs and to explain decisions to farmers or stakeholders.<\/span><\/p><\/li>

                            • Deployment<\/span>: Adapting deep learning models for mobile devices and IoT platforms remains a challenge. Optimizing models for lower computational power and real-time applications while maintaining accuracy is an ongoing area of research.<\/span><\/p><\/li><\/ul>

                              \u00a0<\/b><\/p>

                              Future Directions<\/span><\/p>

                              While deep learning has shown great promise in agricultural disease detection, several emerging research directions can significantly advance this field.<\/span><\/p>

                              • Self-Supervised Learning <\/span>: To address the challenge of limited labeled data, self-supervised learning enables models to learn useful representations from large volumes of unlabeled images by solving pretext tasks. This can help improve performance when labeled data for rare or emerging diseases is scarce.<\/span><\/p><\/li>

                              • Few-Shot Learning<\/span>: Few-shot learning aims to train models that can recognize new disease classes using only a small number of labeled samples. This approach is critical in agricultural settings, where new diseases may appear unexpectedly, and large annotated datasets are unavailable.<\/span><\/p><\/li>

                              • Explainable AI (XAI)<\/span>: Increasing the transparency and interpretability of deep learning models is essential for building trust among farmers and agricultural stakeholders. XAI techniques can help visualize which image regions influenced the model\u2019s prediction, making decisions more understandable and actionable.<\/span><\/p><\/li>

                              • Edge Computing and IoT Integration<\/span>: Deploying lightweight deep learning models on mobile devices, drones, and IoT sensors enables real-time, on-field disease detection without the need for constant internet connectivity. This will be particularly valuable in rural and resource-limited areas.<\/span><\/p><\/li>

                              • Multimodal Fusion<\/span>: Combining image data with other sensor modalities, such as temperature, humidity, and soil conditions, can improve the accuracy and robustness of disease detection systems. Multimodal models can better capture the complex interactions between environmental factors and plant health, leading to more holistic agricultural decision support. Together, these directions will drive the development of next-generation precision agriculture tools, improving resilience and sustainability in farming systems.<\/span><\/p><\/li><\/ul>

                                \"C:\\Users\\tmu\\Downloads\\c7c4ca9d-4e60-4288-844a-a69296304502.jpg\"<\/span><\/span><\/p>

                                Figure 4: Future Framework Integrating Sensors, DL, and IoT<\/span><\/p>

                                Conclusion<\/span><\/p>

                                Deep learning has revolutionized the field of agricultural disease detection, providing rapid, scalable, and precise solutions that were unimaginable just a decade ago. By leveraging the power of Convolutional Neural Networks (CNNs), transformer-based models, and hybrid architectures, researchers have achieved outstanding performance on benchmark datasets, demonstrating the capability to accurately classify a wide variety of plant diseases from image data. These advances hold tremendous potential for improving agricultural productivity, reducing pesticide use, and enhancing global food security.<\/span><\/p>

                                Despite these promising results, several critical challenges must be addressed to enable the practical deployment of deep learning systems in real-world agricultural settings. Data scarcity remains a major limitation, particularly for rare or newly emerging diseases where labeled datasets are limited. Furthermore, models trained on controlled datasets often struggle to generalize to field conditions due to variations in lighting, background, weather, and plant varieties. Another pressing concern is the lack of explainability, as deep learning models often function as \u201cblack boxes,\u201d making it difficult to interpret their decisions and build trust among farmers and agricultural stakeholders.<\/span><\/p>

                                To overcome these limitations, future research should prioritize the development of robust, explainable, and lightweight deep learning models that can operate effectively under diverse environmental conditions. Emphasis should also be placed on self-supervised and few-shot learning methods to reduce dependence on large labeled datasets. Additionally, integrating deep learning with edge computing and Internet of Things (IoT) devices will allow real-time disease detection in the field, even in resource-constrained regions. By addressing these challenges, the next generation of agricultural AI tools will empower farmers worldwide, leading to more resilient, sustainable, and productive farming systems.<\/span><\/p>

                                \u00a0<\/b><\/p>

                                References:<\/span><\/p>

                                1. Alam, M., Verma, P., & Singh, L. (2025). Edge-AI systems for real-time plant disease monitoring. <\/span>IEEE Internet of Things Journal<\/span>, <\/span>9<\/span>(3), 1234\u20131245.<\/span><\/p><\/li>

                                2. Challenger, A. I. <\/span>Agriculture Dataset<\/span> [Online]. <\/span>https:\/\/challenger.ai<\/span><\/p><\/li>

                                3. Chen, X., Fan, H., Girshick, R., & He, K. (2020). Improved baselines with momentum contrastive learning. In <\/span>ICML<\/span>.<\/span><\/p><\/li>

                                4. Dosovitskiy, A. et al. (2021). An image is worth 16\u00d716 words: Transformers for image recognition at scale. In <\/span>ICLR<\/span>.<\/span><\/p><\/li>

                                5. Gupta, N., & Banerjee, S. (2024). Few-shot learning in precision agriculture: A case study on wheat rust detection. <\/span>Computers and Electronics in Agriculture<\/span>, <\/span>215<\/span>, Article 106890.<\/span><\/p><\/li>

                                6. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In CVPR.<\/span><\/p><\/li>

                                7. Kaggle Plant Disease Detection Dataset<\/span> [Online]. <\/span>https:\/\/www.kaggle.com<\/span><\/p><\/li>

                                8. Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. In <\/span>NeurIPS<\/span>, 1106\u20131114.<\/span><\/p><\/li>

                                9. Kumar, A., Patel, S., & Sharma, R. (2024). A transformer-based model for multiclass plant disease classification. <\/span>IEEE Transactions on Artificial Intelligence<\/span>, <\/span>5<\/span>(2), 180\u2013192.<\/span><\/p><\/li>

                                10. Liu, J., Lee, K., & Rahman, M. (2024). Self-supervised learning for agricultural image analysis. In <\/span>Proceedings of the CVPR Workshops<\/span> (pp. 1450\u20131459).<\/span><\/p><\/li>

                                11. Mohanty, M., Hughes, D. P., & Salath\u00e9, M. (2018). Image-based plant disease detection: A review. <\/span>Computers and Electronics in Agriculture<\/span>, <\/span>144<\/span>, 118\u2013132.<\/span><\/p><\/li>

                                12. Picon, D., Ceballos, M., & Garcia, J. (2019). Deep learning applications in agriculture. <\/span>Agronomy<\/span>, <\/span>9<\/span>, 224\u2013238.<\/span><\/p><\/li>

                                13. PlantVillage Dataset<\/span> [Online]. <\/span>https:\/\/plantvillage.psu.edu<\/span><\/p><\/li>

                                14. Qiu, T., Chen, N., Li, K., & Min, G. (2020). Edge computing for agricultural IoT: Architectures, applications, and challenges. <\/span>IEEE Internet of Things Journal<\/span>, <\/span>7<\/span>(5), 4221\u20134230.<\/span><\/p><\/li>

                                15. Simonyan, K., & Zisserman, A. (2014). <\/span>\u2018Very Deep Convolutional Networks for Large-Scale Image Recognition,\u2019 arXiv preprint arXiv:1409.1556<\/span>.<\/span><\/p><\/li>

                                16. Tan, M., & Le, Q. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In <\/span>ICML<\/span>.<\/span><\/p><\/li>

                                17. Torres, F., Singh, A., & Mehta, K. (2025). Explainable Deep learning for disease prediction in crops. <\/span>Artificial Intelligence in Agriculture<\/span>, <\/span>9<\/span>, 45\u201357.<\/span><\/p><\/li>

                                18. Wang, X., Girshick, R., Gupta, A., & He, K. (2018). Non-local neural networks. In <\/span>CVPR<\/span><\/p><\/li>

                                19. Zhang, P., Li, J., Wang, Y., & Liu, H. (2020). Global agricultural losses. <\/span>Nature Food<\/span>, <\/span>1<\/span>(2), 123\u2013129.<\/span><\/p><\/li>

                                20. Zhang, S. W., Huang, X. Z., & Zhang, Y. S. (2015). Plant disease recognition based on KNN. <\/span>Computer Engineering and Science<\/span>, <\/span>37<\/span>, 184\u2013188.<\/span><\/p><\/li>

                                21. Zhang, Z., Chen, Y., & Zhang, L. (2021). Multimodal data fusion for precision agriculture. <\/span>Remote Sensing<\/span>, <\/span>13<\/span>(9).<\/span><\/p><\/li><\/ol>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

                                  \n\t\t\t\t\t\t
                                  \n\t\t\t\t\t
                                  \n\t\t\t
                                  \n\t\t\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"

                                  Disease Detection in Agriculture using Deep Learning Abhishek Sharma Research Scholar, Teerthanker Mahaveer University, Moradabad Disease Detection in Agriculture using Deep Learning \u00a0 Abhishek Sharma Research Scholar, Teerthanker Mahaveer University, Moradabad Abstract Agriculture is a fundamental pillar of global food security and economic stability. However, plant diseases pose a severe threat to agricultural productivity, leading […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":13068,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"site-sidebar-layout":"no-sidebar","site-content-layout":"page-builder","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"disabled","ast-breadcrumbs-content":"","ast-featured-img":"disabled","footer-sml-layout":"","theme-transparent-header-meta":"default","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"class_list":["post-13205","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/pages\/13205","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/comments?post=13205"}],"version-history":[{"count":19,"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/pages\/13205\/revisions"}],"predecessor-version":[{"id":13399,"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/pages\/13205\/revisions\/13399"}],"up":[{"embeddable":true,"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/pages\/13068"}],"wp:attachment":[{"href":"https:\/\/icertpublication.com\/index.php\/wp-json\/wp\/v2\/media?parent=13205"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}