The current study outlines an overall machine-learning model of the automated detection of disease-inducing factors in sugarcane foliage, using hybrid, image-based feature engineering coupled with more superior models of classification. A total of 2,521 images of five major disease groups, such as Healthy, Mosaic, Red Rot, Rust and Yellow Leaf Disease, were taken through preprocessing processes which involved resizing, denoising and translation into HSV colour space to enable increased chromatic languages. A hybrid feature-extraction approach, which involves the combination of classical feature descriptions HOG and deep convolutional features, which were achieved through EfficientNetB0, was used. These multidimensional hybrid features were further reduced to a 300-dimensional discriminative vector via the Principal Component Analysis hence significantly lowering the computational costs without losing any critical information. Several nearest classifiers, such as Random Forest and XGBoost were trained on the smaller set of features. The results of comparative performance analysis showed that XGBoost was more successful in performance as it reached a level of accuracy of 83.82 as opposed to the accuracy of 56.21 in the case of the Random Forest model. The consistency of the XGBoost classifier to separate disease categories that were similar was verified by further analysis with precision metrics, recall metrics, F1-score, and confusion matrices. The suggested system can be proven to be robust, efficient and scalable in detecting early disease in sugarcane crops.

The paper presents a complete and systematic system of the automated detection of sugarcane diseases through machine learning using images. To test and train the proposed system, a collection of 2,521 leaf images representing five main types of diseases, including Mosaic, Red Rot, Rust, Yellow Leaf Disease, and Healthy, was used. The process will involve necessary steps such as image preprocessing, classical features creation and convolutional neural network (CNN) with deep features generation, hybrid feature fusion, dimensionality reduction via principal components analysis (PCA) and machine-learning-based classification. The classical module of feature obtained discriminative colour, texture and structural features. These complementary features were summed together to create a hybrid feature and PCA was used to reduce the size of the hybrid feature to 300 dimensions, keeping computation methods small and curtailing overfitting. Random Forest and XGBoost were tested as two classifiers using reduced feature set. The Random Forest was only able to obtain a moderate accuracy of 56.21%, which indicates that the problem of characterizing the complex feature space is difficult. Conversely, XGBoost achieved an accuracy of 83.82% and it has a better ability to model non-linearized relationship and to model high-dimensional integrated features. The incidence of improvement of the precision, recall, and F1-score support the claim that XGBoost is a more robust and dependable classifier of sugarcane leaf disease.

[1] Reddy, A. V., Thiruvengatanadhan, R., Srinivas, M., & Dhanalakshmi, P. (2023). Artificia l intelligence framework for sugarcane diseases classification using convolutiona l neural network. Int. J. Recent Innov. Trends Comput. Commun, 11(9), 3620-3628. [2] Akhil, S., & Durga, R. (2025, June). Analysis and Detection of Diseases in Leaf Images for Sugarcane using Efficient PARNET-52 Deep Learning Techniques. In The 2025 International Conference on Advanced Research in Electronics and Communication Systems (ICARECS-2025) (pp. 121-133). Atlantis Press. [3] Vivekreddy, A., Thiruvengatanadhan, R., Srinivas, M., & Dhanalakshmi, P. (2024). Artificial Intelligence Framework for Multi-Class Sugarcane Leaf Diseases Classification Using Deep Learning Algorithms. Journal of Theoretical and Applied Information Technology, 31(10). [4] Bao, D., Zhou, J., Bhuiyan, S. A., Adhikari, P., Tuxworth, G., Ford, R., & Gao, Y. (2024). Early detection of sugarcane smut and mosaic diseases via hyperspectral imaging and spectral-spatial attention deep neural networks. Journal of Agriculture and Food Research, 18, 101369. [5] Nomani, K. M. S., Nuruzzaman, M., Nadia, A., & Billal, M. M. (2024, October). Sugarcane Leaf Disease Detection: A Comparative Analysis Using Deep Learning. In Proceedings of the 3rd International Conference on Computing Advancements (pp. 139-144). [6] Islam, M. M., Adil, M. A. A., Talukder, M. A., Ahamed, M. K. U., Uddin, M. A., Hasan, M. K., … & Debnath, S. K. (2023). Deep Crop: Deep learning-based crop disease prediction with web application. Journal of Agriculture and Food Research, 14, 100764. [7] Nomani, K. M. S., Nuruzzaman, M., Nadia, A., & Billal, M. M. (2024, October). Sugarcane Leaf Disease Detection: A Comparative Analysis Using Deep Learning. In Proceedings of the 3rd International Conference on Computing Advancements (pp. 139-144). [8] Wadhe, V., Dongre, R., Kankriya, Y., & Kuckian, A. (2022, December). Sugarcane disease detection using deep learning. In 2022 5th International Conference on Advances in Science and Technology (ICAST) (pp. 40-44). IEEE. [9] Upadhye, S. A., Dhanvijay, M. R., & Patil, S. M. (2022, December). Sugarcane disease detection using CNN-deep learning method. In 2022 Fourth International Conference on Emerging Research in Electronics, Computer Science and Technology (ICERECT) (pp. 1-8). IEEE. [10] Santhrupth, B. C., & Devaraj Verma, C. (2024). Intelligent disease detection in sugarcane plants: A comparative analysis of machine learning models for classification and diagnosis. International Journal of Intelligent Systems and Applications in Engineering, 12(8s), 299-306. [11] Ngugi, H.N., Akinyelu, A. A., & Ezugwu, A. E. (2024). Machine learning and deep learning for crop disease diagnosis: Performance analysis and review. Agronomy, 14(12), 3001. [12] Shoaib, M., Shah, B., Ei-Sappagh, S., Ali, A., Ullah, A., Alenezi, F., … & Ali, F. (2023). An advanced deep learning models-based plant disease detection: A review of recent research. Frontiers in Plant Science, 14, 1158933. [13] Mathew, A. P., Viswajith, P., Rahman, A., & Murali, V. (2023). Crop disease detection using machine learning. Journal of Applied Science, Engineering, Technology and Management, 1(01), 28-32. [14] https://www.kaggle.com/datasets/pritpal2873/sugarcane-leaf-disease-dataset