The Indian economy relies heavily on agriculture, and rice is one of the most significant staple crops grown in the country. Nevertheless, rice cultivation faces numerous challenges due to several leaf diseases, such as bacterial leaf blight, brown spot, and leaf smut. These diseases can lead to substantial yield losses and financial burdens for farmers. Therefore, the early detection and classification of these diseases are essential to minimize crop damage and improve crop management practices. Conventional approaches to disease detection rely on visual observations by experts. However, this method is time-consuming, costly, and prone to human error, especially when symptoms are difficult to observe or appear similar to one another.
To address this problem, the proposed project develops an image-based rice crop disease detection system using deep learning and Grad-CAM for accurate disease identification. The dataset was collected from Kaggle and contains more than 18,694 images of rice leaves categorized into different disease classes as well as healthy leaves. Data preprocessing techniques were applied to balance the dataset and improve its robustness, including resizing, normalization, and augmentation.
A Convolutional Neural Network (CNN) was implemented for automatic feature extraction, while the advanced object detection model YOLOv8 was employed to localize diseased regions with high speed and accuracy. Additionally, Grad-CAM (Gradient-weighted Class Activation Mapping) was integrated to provide visual interpretability of the model by highlighting the regions that significantly influenced its predictions. Experimental results demonstrated a classification accuracy of 86.41%, on the test set, outperforming traditional machine learning algorithms and earlier deep neural network architectures.

This study proves the significance of deep learning in Qualitative farming. It also points out that explainable AI is necessary in increasing the trust between the farmers and the agricultural professionals. The model proposed provides a scaled, resourceful and transparent approach to real time detection of rice disease. Other studies such as other crops can adopt this method. Finally, this project should contribute to the general objective of sustainable agriculture, which is the integration of artificial intelligence with image analysis.

Keywords—Rice disease detection, Deep learning, CNN, YOLOv8, Grad-CAM, Explainable AI, Agriculture

🔒 This is a subscription article

Full text is available to subscribers and institutional members. Please choose an option below to access it.

• Subbarayudu C, Kubendiran M. An automated hybrid deep learning framework for paddy leaf disease identification and classification. Scientific Reports. 2025 Jul 24;15(1):26873.
• Naresh Kumar B, Sakthivel S. Rice leaf disease classification using a fusion vision approach. Scientific Reports. 2025 Mar 13;15(1):8692.
• Cao Q, Zhao D, Li J, Li J, Li G, Feng S, Xu T. Pyramid-YOLOv8: a detection algorithm for precise detection of rice leaf blast. Plant methods. 2024 Sep 28;20(1):149.
• Gan B, Pu G, Xing W, Wang L, Liang S. Enhanced YOLOv8 with lightweight and efficient detection head for for detecting rice leaf diseases. Scientific Reports. 2025 Jul 1;15(1):22179.
• Thai-Nghe N, Tri NT, Hoa NH. Deep learning for rice leaf disease detection in smart agriculture. InInternational Conference on Artificial Intelligence and Big Data in Digital Era 2021 Dec 18 (pp. 659-670). Cham: Springer International Publishing.
• Petchiammal, Kiruba B, Murugan, Arjunan P. Paddy doctor: A visual image dataset for automated paddy disease classification and benchmarking. InProceedings of the 6th Joint International Conference on Data Science & Management of Data (10th ACM IKDD CODS and 28th COMAD) 2023 Jan 4 (pp. 203-207).
• Murugan D. Paddy Leaf diseases identification on Infrared Images based on Convolutional Neural Networks. arXiv preprint arXiv:2208.00031. 2022 Jul 29.

The Indian economy relies heavily on agriculture, and rice is one of the most significant staple crops grown in the country. Nevertheless, rice cultivation faces numerous challenges due to several leaf diseases, such as bacterial leaf blight, brown spot, and leaf smut. These diseases can lead to substantial yield losses and financial burdens for farmers. Therefore, the early detection and classification of these diseases are essential to minimize crop damage and improve crop management practices. Conventional approaches to disease detection rely on visual observations by experts. However, this method is time-consuming, costly, and prone to human error, especially when symptoms are difficult to observe or appear similar to one another.
To address this problem, the proposed project develops an image-based rice crop disease detection system using deep learning and Grad-CAM for accurate disease identification. The dataset was collected from Kaggle and contains more than 18,694 images of rice leaves categorized into different disease classes as well as healthy leaves. Data preprocessing techniques were applied to balance the dataset and improve its robustness, including resizing, normalization, and augmentation.
A Convolutional Neural Network (CNN) was implemented for automatic feature extraction, while the advanced object detection model YOLOv8 was employed to localize diseased regions with high speed and accuracy. Additionally, Grad-CAM (Gradient-weighted Class Activation Mapping) was integrated to provide visual interpretability of the model by highlighting the regions that significantly influenced its predictions. Experimental results demonstrated a classification accuracy of 86.41%, on the test set, outperforming traditional machine learning algorithms and earlier deep neural network architectures.

This study proves the significance of deep learning in Qualitative farming. It also points out that explainable AI is necessary in increasing the trust between the farmers and the agricultural professionals. The model proposed provides a scaled, resourceful and transparent approach to real time detection of rice disease. Other studies such as other crops can adopt this method. Finally, this project should contribute to the general objective of sustainable agriculture, which is the integration of artificial intelligence with image analysis.

Keywords—Rice disease detection, Deep learning, CNN, YOLOv8, Grad-CAM, Explainable AI, Agriculture

🔒 This is a subscription article

Full text is available to subscribers and institutional members. Please choose an option below to access it.

• Subbarayudu C, Kubendiran M. An automated hybrid deep learning framework for paddy leaf disease identification and classification. Scientific Reports. 2025 Jul 24;15(1):26873.
• Naresh Kumar B, Sakthivel S. Rice leaf disease classification using a fusion vision approach. Scientific Reports. 2025 Mar 13;15(1):8692.
• Cao Q, Zhao D, Li J, Li J, Li G, Feng S, Xu T. Pyramid-YOLOv8: a detection algorithm for precise detection of rice leaf blast. Plant methods. 2024 Sep 28;20(1):149.
• Gan B, Pu G, Xing W, Wang L, Liang S. Enhanced YOLOv8 with lightweight and efficient detection head for for detecting rice leaf diseases. Scientific Reports. 2025 Jul 1;15(1):22179.
• Thai-Nghe N, Tri NT, Hoa NH. Deep learning for rice leaf disease detection in smart agriculture. InInternational Conference on Artificial Intelligence and Big Data in Digital Era 2021 Dec 18 (pp. 659-670). Cham: Springer International Publishing.
• Petchiammal, Kiruba B, Murugan, Arjunan P. Paddy doctor: A visual image dataset for automated paddy disease classification and benchmarking. InProceedings of the 6th Joint International Conference on Data Science & Management of Data (10th ACM IKDD CODS and 28th COMAD) 2023 Jan 4 (pp. 203-207).
• Murugan D. Paddy Leaf diseases identification on Infrared Images based on Convolutional Neural Networks. arXiv preprint arXiv:2208.00031. 2022 Jul 29.