Banana Leaf Disease Detection and Classification Using CNN Model

One of the most extensively farmed tropical fruits is the banana. It produces 16% of global fruit, second only to citrus. Leaves diseases affect banana crop production which results in poor economic condition of banana farmers. This research focuses on the detection and classification of banana plant diseases using deep learning and transfer learning models. The primary objective is to develop a reliable tool for farmers which enables early and accurate banana plant disease detection. Researchers employed multiple convolutional neural network models, such as VGG19, LeNet, ResNet101V2, MobileNetV2, and InceptionV3, to classify six different banana plant diseases, namely Insect Pests, Cordana, Pestalotiopsis, Sigatoka, Bract Mosaic, and Moko. The data has undergone appropriate data processing steps, and each model was fine-tuned for optimized results. VGG19 model has delineated the best performance, achieving an accuracy of 97%. This study highlights the potential of transfer learning models used in plant disease detection tasks and provides a scalable solution for managing crop health.