Pea Plant Leaf Disease Detection Using ResNetV2

Abstract

Early and accurate detection of pea plant leaf diseases is critical for improving crop yield and preventing widespread damage. This paper presents a deep learning-based approach for automated classification of pea leaf diseases using the ResNetV2 convolutional neural network architecture. The study utilizes a publicly available pea plant leaf dataset comprising four classes: Downy Mildew, Powdery Mildew, Leafminer damage, and Healthy leaves. The dataset was preprocessed and augmented to enhance model generalization, and the ResNetV2 model was fine-tuned to achieve effective feature extraction and classification. Experimental results demonstrate that the proposed method achieves a validation accuracy of approximately 94%, outperforming baseline models. The model’s performance was further analyzed via precision, recall, F1-score, and confusion matrix, confirming its robustness across all disease categories. The findings indicate that ResNetV2 is a promising candidate for practical deployment in agricultural monitoring systems, enabling timely disease diagnosis and management.

Introduction

Pea plants (Pisum sativum) are globally important crops whose productivity is threatened by leaf diseases like Downy Mildew, Powdery Mildew, and Leafminer damage. Traditional manual disease detection is slow, laborious, and error-prone, especially for large-scale farming. Recent advances in deep learning, particularly convolutional neural networks (CNNs), offer automated, accurate, and scalable disease classification from leaf images.

This study uses the ResNetV2 CNN architecture, pretrained on ImageNet, to classify pea leaf images into four categories: Downy Mildew, Powdery Mildew, Leafminer damage, and Healthy leaves. A dataset of 1,432 RGB images was preprocessed (resized, normalized) and augmented with rotations, flips, zooms, and shifts to improve model robustness. The model was trained in two stages—first freezing base layers, then fine-tuning the last 20 layers—with categorical cross-entropy loss and Adam optimizer.

Evaluation on a validation set of 285 images yielded an overall accuracy of 94%, with F1-scores above 89% for all classes. The model performed best in identifying Healthy and Powdery Mildew leaves, while Downy Mildew had slightly lower recall due to visual similarities causing some misclassification. The confusion matrix confirmed strong classification performance with few errors.

The study concludes that ResNetV2 is effective for automated pea leaf disease detection, offering a promising tool for real-world agricultural monitoring, with data augmentation crucial in preventing overfitting.

Conclusion

In summary, this work demonstrates that a ResNetV2-based deep learning model can achieve high accuracy for multi-class pea plant leaf disease detection. Advanced data augmentation and careful training strategies contributed to robust generalization and reliable performance, confirming the value of modern CNN architectures for automated plant pathology. For future research, expanding the dataset to include additional disease classes and more field samples would likely improve model resilience and accuracy. Implementing the system for real-time diagnosis on mobile or edge devices could support practical adoption among farmers and extension specialists. Further, exploring techniques for improved explainability and user-friendly interfaces will enhance the impact of such solutions in precision agriculture.

References

[1] K. He, X. Zhang, S. Ren, and J. Sun, “Identity Mappings in Deep Residual Networks,” in European Conference on Computer Vision (ECCV), 2016, pp. 630–645. [2] K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770–778. [3] S. Mohanty, D. P. Hughes, and M. Salathé, “Using Deep Learning for Image-Based Plant Disease Detection,” Frontiers in Plant Science, vol. 7, 2016. [4] S. Sladojevic, M. Arsenovic, A. Anderla, D. Culibrk, and D. Stefanovic, “Deep Neural Networks Based Recognition of Plant Diseases by Leaf Image Classification,” Computational Intelligence and Neuroscience, vol. 2016, Article ID 3289801. [5] M. Ali, R. Kumar, “Field Pea Leaf Disease Classification using a Deep Learning Model Trained on a Small, Imbalanced Dataset,” PLOS ONE, vol. 19, no. 7, 2024. [6] Zunorain, Abdullah, “Pea Plant Leaf Disease Detection Dataset,” figshare, Dataset, 2025. https://doi.org/10.6084/m9.figshare.29254046. [7] A. Kamilaris and F. Prenafeta-Boldú, “Deep Learning in Agriculture: A Survey,” Computers and Electronics in Agriculture, vol. 147, pp. 70– 90, 2018. [8] A. Ferentinos, “Deep Learning Models for Plant Disease Detection and Diagnosis,” Computers and Electronics in Agriculture, vol. 145, pp. 311–318, 2018. [9] T. Brahimi, K. Boukhalfa, and A. Moussaoui, “Deep Learning for Tomato Diseases: Classification and Symptoms Visualization,” Applied Artificial Intelligence, vol. 31, no. 4, pp. 299–315, 2017. [10] S. Arsenovic, S. Sladojevic, A. Anderla, and D. Culibrk, “A Deep Learning-Based Approach for the Detection of Diseases in Wheat Leaves,” Computers and Electronics in Agriculture, vol. 158, pp. 97–107, 2019. [11] Y. LeCun, Y. Bengio, and G. Hinton, “Deep Learning,” Nature, vol. 521, pp. 436–444, 2015. [12] S. Kamilaris, F. X. Prenafeta-Boldú, “Detecting Prevalent Crop Diseases Using Deep Learning Algorithms,” International Journal of Agricultural Research, vol. 13, no. 2, pp. 1–12, 2018. [13] Z. Zhang, “Transfer Learning for Plant Disease Identification,” in Proceedings of the International Conference on Computational Science, 2021, pp. 250-256.

Copyright

Copyright © 2025 Aditya Bisht, Aabhas Vishwas, Abhishek Maurya, Abhyuday Pundir. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.