A Real-Time Application-Based Bidirectional Long Short-Term Memory-Based Adam Optimizer for Leaf Disease Classification in Sugarcane Leaves

Sugarcane is considered to be a vital crop across worldwide, but the production of it is impacted by the effect of different diseases. Early diagnosis and detection are mandatory for the timely intervention and yield preservation. This will ensure the optimal yield of the crop and its quality. Hence in this approach a real-time hybrid approach BADAM is proposed for the detection and classification of the sugarcane leaf diseases. The hybrid approach BADAM consists of Bidirectional Long Short-Term Memory (BiLSTM) network combined with the Adam optimizer. For training the model both disease-affected leaves and the nonaffected leaves are considered and initially subjected to preprocessing model for enhancing the feature extraction process. The BiLSTM model is designed to capture both temporal and spatial dependencies within the image sequences, effectively addressing the challenges posed by varying leaf conditions and environmental factors. The Adam optimizer is employed to improve convergence rates and enhance model performance through adaptive learning rates. The model’s performance is evaluated using the metrics like accuracy, recall, [Formula: see text]-score and precision by demonstrating significant improvements over traditional classification methods. The experimentation results indicate that the proposed approach provides better accuracy by facilitating the real-time disease detection by making it comfortable by providing a valuable tool for farmers and agricultural practitioners. The implications of this work extend to the development of smart agricultural systems that leverage deep learning technologies for sustainable farming practices. This approach not only provides an effective tool for leaf disease classification but also establishes a foundation for future research in precision agriculture and plant health monitoring, promoting sustainable farming practices through early disease detection and intervention.