Deep Learning-Based Automated Detection of Rice Leaf Diseases Using Image Processing

Abstract

Rice is one of the vital sources of food in the world, but the process of growing it becomes increasingly challenging due to the presence of various diseases in its leaves. Manual detection of such diseases takes a long time and sometimes even leads to misdiagnosis. Thus, we propose a holistic approach to detect diseases on the leaves of rice plants effectively. We employ Convolutional Neural Networks (CNN). In our work, we design a CNN using TensorFlow and Keras for proper image classification. Our solution includes a back-end part written in Python and Flask framework and a front-end part implemented in React.js. We train the model based on the “rice-disease-dataset” to classify healthy rice plants and their leaves that suffer from blast, blight, and brown spots. The achieved accuracy rate reaches 95.60%, thus enabling farmers to diagnose crops timely.

Introduction

RiceCare-AI is an AI-based system designed to automatically detect and classify rice leaf diseases such as Brown Spot, Leaf Blast, Leaf Blight, and healthy leaves. It addresses the limitations of traditional manual diagnosis, which is slow, error-prone, and difficult to access in rural areas. The system uses a Convolutional Neural Network (CNN) built with TensorFlow and Keras, along with image preprocessing, data augmentation, and optimized training techniques to improve accuracy and robustness.

The model is trained on a labeled dataset of rice leaf images collected from agricultural sources and Kaggle, with preprocessing steps like resizing, normalization, and augmentation to handle real-world variations. It achieves high performance (about 95.6% accuracy) and is evaluated using precision, recall, and F1-score. The system is also optimized for real-time use through lightweight deployment using TensorFlow Lite.

RiceCare-AI is deployed as a full-stack application using a Flask backend and React.js frontend, allowing users (farmers or researchers) to upload images and receive instant disease predictions. Despite strong performance, the system is limited to a few disease classes and may struggle under complex field conditions. Future improvements include expanding the dataset, handling more disease types, and using advanced imaging techniques for better accuracy and real-world applicability.

Conclusion

It was found that RiceCare-AI is an efficient and robust fully automated system that can detect and classify diseases on rice leaves. This model could classify the difference between normal leaves and other diseases, such as Brown Spot, Leaf Blast, and Leaf Blight, through the implementation of an enhanced CNN algorithm that was created using TensorFlow and Keras. Additionally, the accuracy of the model was recorded at 95.60%. One of the significant achievements of this research paper was successfully implementing the trained model within a functional full-stack system that comprises both a Flask back-end and a React.js front-end. This integration not only allows for low-latency, real-time inference but also creates an easy-to-use user interface that is applicable to farmers and agriculturalists. From the analysis above, the proposed framework would be reliable to operate in different environments, hence a significant tool that could help detect the disease at its earliest stage. This would be an indispensable asset in facilitating sustainable agricultural production, minimizing crop losses, and implementing effective intervention strategies. Even though the system performs excellently, its scope of application remains restricted to only a few disease categories. In future research, efforts will be made to broaden the data pool to account for various other plant diseases that occur locally and have unique characteristics. Moreover, using high-end imaging technologies like hyperspectral and thermal imaging might prove to be beneficial. To conclude, RiceCare-AI is a revolutionary step toward digital agriculture because it helps improve crop management and ensures food security.

References

[1] Gopi, S. C. and Kondaveeti, H. K. 2023. Automatic recognition of rice leaf diseases using transfer learning. Agronomy, 13(4): 961. [2] Kumar, B. N. and Sakthivel, S. 2025. Rice leaf disease classification using a fusion vision approach. Scientific Reports, 15(1): 87800. [3] Tasci, M. 2025. DBLA-MobileNetV2: Real-time rice leaf disease detection on edge devices with dual-branch lightweight attention. IEEE Access, 13: 45012–45028. [4] Singha Roy, P. and Kukreja, V. 2026. Deep learning-based identification of rice leaf diseases: A systematic review. Journal of Transformative Technologies and Sustainable Development, 10(1). [5] Pai, P. et al. 2024. Deep learning-based automatic diagnosis of rice leaf diseases using ensemble CNN models. Scientific Reports, 14(1). [6] Rahman, K. S. et al. 2024. Assessment of paddy leaves disease severity level using image processing technique. Smart Agricultural Technology, 7: 100410. [7] Hartono, R., Yoeseph, N. M., Purnomo, F. A., Bawono, S. A. T. and Purnomo, A. 2025. A lightweight convolutional neural network for rice leaf disease detection integrated in an Android application. Bulletin of Electrical Engineering and Informatics, 14(4): 2705–2720. [8] Acharya, S. et al. 2026. Automated detection of rice plant diseases using dual stage thresholding and twin support vector machine. Bulletin of Electrical Engineering and Informatics, 15(1): 512–521. [9] Devi, A. G. et al. 2026. Multi-task deep learning framework for segmentation and severity estimation of leaf diseases in multi-crop environments. International Journal of Innovative Technology and Interdisciplinary Sciences, 9(1): 210–237. [10] Rani, A. P. A. S. 2026. An efficient hybrid convolutional vision transformer framework with spatial attention for rice leaf disease identification and categorization. Research Square.

Copyright

Copyright © 2026 Vaibhav R. Bhadane, Kuldip B. Mali, Nandlal V. Rajput, Zuber Shaikh, Prof. Santosh Kumar Bhandare. 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.