Sugarcane Disease Classification using Advanced Deep Learning Algorithms: A Comparative Study

Accurately identifying sugarcane diseases is crucial for farmers. This study explores how advanced deep learning models can be used for this task. We compared different models and created a new hybrid model that combines the strengths of EfficientNetBO, a custom-designed neural network, and CSPDarknet53. The custom model is designed to capture the unique features of sugarcane diseases. By combining the information learned by these different models, our hybrid model achieves a remarkable accuracy of 96.8 % on a standard dataset. This research shows that combining different deep learning architectures can lead to superior performance in disease clas-sification tasks. This approach not only achieves high accuracy but also offers a robust and adaptable solution for real-world agricultural applications. Our findings contribute to the development of reliable deep learning tools for crop health monitoring and precision agriculture.