Soybean Leaf Disease Detection System Using Convolutional Neural Network

Abstract

Soybean cultivation contributes significantly to global food security and industrial supply chains. However, soybean leaf diseases significantly affect crop yield and quality, and traditional manual disease identification methods are often slow, error-prone, and dependent on expert knowledge to be effective. In this study, a Convolutional Neural Network (CNN)-based automated soybean leaf disease detection system was developed. The model was trained on a dataset containing six disease classes using image preprocessing, augmentation, and classification techniques. The proposed system achieved a training accuracy of 95.79% and a validation accuracy of 84.06%, demonstrating strong generalization on unseen images. A web-based interface was implemented using Flask to provide real-time image uploads and disease prediction. The proposed system is efficient, scalable, and beneficial for farmers to detect diseases early, reduce crop losses, and promote smart agriculture.

Introduction

Soybean is an important crop used for food, livestock feed, and industrial purposes because of its high protein and oil content. However, soybean plants are often affected by leaf diseases such as Bacterial Blight, Brown Spot, Powdery Mildew, Rust, and Yellow Mosaic Virus (YMV), which reduce crop yield and quality. Traditional disease detection methods rely on visual inspection by farmers, which can be inaccurate due to similar disease symptoms and lack of expert guidance, especially in rural areas.

To address this problem, the study proposes an automated soybean leaf disease detection system using deep learning, particularly Convolutional Neural Networks (CNNs). The system was trained on a dataset of 1392 soybean leaf images obtained from Kaggle, including 1116 training images and 276 validation images. Image preprocessing and data augmentation techniques were applied to improve model performance.

Two models were tested: a baseline CNN model and a MobileNetV2 transfer-learning model, with MobileNetV2 showing better performance. The proposed model achieved 95.79% training accuracy and 84.06% validation accuracy, demonstrating strong capability in identifying different soybean diseases. The system also used performance metrics such as precision, recall, and F1-score to evaluate classification results.

The results showed that the model effectively recognized disease patterns such as leaf discoloration, lesions, fungal layers, and rust spots. However, some misclassifications occurred between diseases with similar visual patterns, such as Brown Spot and Yellow Mosaic. Overall accuracy from the confusion matrix was 76%, which is considered acceptable for real-world agricultural datasets.

The developed system demonstrates that deep learning-based image analysis can support early disease detection, reduce reliance on manual inspection, minimize unnecessary pesticide use, and improve crop management. Such systems can be integrated into web or mobile applications, enabling farmers to upload leaf images and receive instant disease predictions, supporting smart and technology-driven agriculture.

Conclusion

The proposed soybean leaf disease detection system clearly demonstrates the strong potential of deep learning technology for solving real-world agricultural problems. By utilizing a Convolutional Neural Network (CNN) along with a transfer-learning-based MobileNetV2 architecture, the model successfully learned important visual features from soybean leaf images and classified six major classes: Bacterial Blight, Brown Spot, Rust, Powdery Mildew, Healthy, and Yellow Mosaic. With proper image preprocessing and augmentation, the system achieved a training accuracy of 95.79%, validation accuracy of 84.06%, and overall test accuracy of 76% based on the confusion matrix. These results confirm that the proposed model can identify most soybean leaf diseases with good reliability under practical conditions. Although some confusion was observed between visually similar diseases, such as Brown Spot and Yellow Mosaic, the overall precision, recall, and F1-score values indicate that the system performs effectively for real-world disease diagnosis.

References

[1] R. J. Barbedo, “Digital image processing techniques for detecting, quantifying and classifying plant diseases,” SpringerPlus, vol. 2, no. 1, pp. 1–12, 2013. [2] A. Pydipati, T. F. Burks, and W. S. Lee, “Identification of citrus disease using color texture features and discriminant analysis,” Computers and Electronics in Agriculture, vol. 52, no. 1–2, pp. 49–59, 2005. [3] A. Camargo and J. S. Smith, “Image pattern classification for the identification of disease-causing agents in plants,” Computers and Electronics in Agriculture, vol. 66, no. 2, pp. 121–125, 2009. [4] A. Camargo and J. S. Smith, “An image-processing based algorithm to automatically identify plant disease visual symptoms,” Biosystems Engineering, vol. 102, no. 1, pp. 9–21, 2009. [5] S. Al-Hiary, S. Bani-Ahmad, M. Reyalat, M. Braik, and Z. ALRahamneh, “Fast and accurate detection and classification of plant diseases,” International Journal of Computer Applications, vol. 17, no. 1, pp. 31–38, 2011. [6] A. Atabay, “Plant leaf classification using deep convolutional neural networks,” in Proc. IEEE Int. Conf. Image Processing, 2016. [7] D. H. Hubel and T. N. Wiesel, “Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex,” Journal of Physiology, vol. 160, no. 1, pp. 106–154, 1962. [8] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998. [9] S. Wu, F. Bao, and E. Xu, “A leaf recognition algorithm for plant classification using probabilistic neural network,” in Proc. IEEE Int. Conf. Computational Science, pp. 449–453, 2007. [10] M. Söderkvist, Computer Vision Classification of Leaves from Swedish Trees, Linköping University, Sweden, 2001. [11] A. K. Reyes, J. C. Caicedo, and J. E. Camargo, “Fine-tuning deep convolutional networks for plant recognition,” in CLEF 2015 Working Notes Papers, Toulouse, France, pp. 1–14, 2015. [12] S. P. Mohanty, D. P. Hughes, and M. Salathé, “Using deep learning for image-based plant disease detection,” Frontiers in Plant Science, vol. 7, pp. 1419, 2016. [13] A. Krizhevsky, I. Sutskever, and G. Hinton, “ImageNet classification with deep convolutional neural networks,” in Advances in Neural Information Processing Systems, vol. 25, pp. 1097–1105, 2012. [14] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted residuals and linear bottlenecks,” in Proc. IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR), pp. 4510–4520, 2018. [15] Kaggle, “Soybean Diseased Leaf Dataset,” Available: https://www.kaggle.com/datasets/sivm205/soybean-diseased-leaf-dataset [16] F. Chollet, Deep Learning with Python, Manning Publications, 2017. [17] J. Amara, B. Bouaziz, and A. Algergawy, “A deep learning-based approach for banana leaf disease classification,” Journal of King Saud University – Computer and Information Sciences, vol. 33, no. 10, pp. 1125–1134, 2021. [18] P. Ferentinos, “Deep learning models for plant disease detection and diagnosis,” Computers and Electronics in Agriculture, vol. 145, pp. 311–318, 2018. [19] H. Jiang, Y. Zhang, M. Qiao, and Y. Chen, “A CNN-based approach for detecting rice leaf diseases,” IEEE Access, vol. 7, pp. 16207–16214, 2019. [20] 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, pp. 1–11, 2016.

Copyright

Copyright © 2026 Rushikesh Sanjay Chandane, Bharti W. Gawali. 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.