Integrated Deep Learning and Optimization Strategies for Accurate Cinnamon Bark Disease Classification

Plant disease identification is crucial for agricultural productivity, especially in crops like cinnamon, which is globally valued for its economic and culinary importance. Diseases can severely affect cinnamon plants, leading to decreased yield and quality. This research aims to automate disease diagnosis and classification in cinnamon plants using deep learning techniques to ensure sustainable cultivation and safeguard this valuable commodity. Multiple deep learning algorithms were applied to a dataset specific to cinnamon species, including ResNet101, GoogleNet, DenseNet, VGG16, and VGG19. Data augmentation techniques were employed to enhance model performance, which resulted in improved accuracy, precision, recall, and F1-score metrics. We proposed Optimized Convolutional Neural Network (OCNN) architectures to further optimize model performance, integrating Elephant Herding Optimization (EHO) with selected CNN architectures. This approach minimizes computation overhead while increasing accuracy by tuning hyperparameters like learning rate, number of epochs, and batch size, enhancing the reliability and effectiveness of disease classification in the automated diagnostic system.