Date Palm White Scale Disease Detection using Convolutional Neural Networks (VGG16)

Date palm (Phoenix dactylifera) is a significant agricultural crop, but its productivity is often hampered by various diseases such as brown spot and white scale infestations. Early detection of these diseases is crucial for effective management and mitigation strategies. In this study, we propose a robust framework for date palm disease detection with help of deep neural network algorithms. The dataset comprising images of healthy date palms, brown spot affected palms, and white scale infested palms is obtained from Kaggle. Prior to model training, the images undergo preprocessing techniques, including median filtering, to enhance their quality and reduce noise. Features are then extracted from the preprocessed images. Two state-of-the-art deep learning models, Convolutional Neural Network (CNN) Inception V3 and VGG-16, are employed for disease classification. The conduct of each model is evaluated using test loss and test accuracy metrics. Our experimental results demonstrate promising performance, with CNN achieving a test loss of 1.29 and a test accuracy of 0.33, Inception V3 achieving a test loss of 0.087 and a test accuracy of 0.93, and VGG-16 achieving a test loss of 0.049 and a test accuracy of 0.98. The comparative analysis reveals that VGG-16 outperforms the remaining models in terms of both accuracy and loss metrics, indicating its efficiency for date palm disease detection tasks. Overall, this study showcases the potential of deep neural networks in automated and effective diagnosis of date palm diseases, thereby facilitating timely interventions and enhancing crop yield.