WOAGE-MobileNetV2 Model for Plantain Tree Pest Classification and Recognition

Plantain trees are one of the most important crops in terms of economic, cultural, and nutritional value. The year-round growth and availability of bananas provide consistent employment and income for the farmers. Plantain trees are highly affected by various pests and diseases. This will impact the health of the plants and overall production systems. Pests are the root cause of major diseases. Early identification of pests helps to control their impact on the health of plantain trees. Deep learning techniques, especially Convolutional Neural Networks (CNNs), have already established their proficiency in the automatic detection of diseases and pests in humans and plants. This research work proposes a CNN model named WOAGE-MobileNetV2, which embeds an enhanced whale optimization algorithm (WOA) with the traditional MobileNetV2 model for improving efficient image classification. The experiments are carried out on the banana pest image dataset for detecting the earwigs, thrips, and weevils. The banana pest image dataset is a real-time dataset collected from Tamil Nadu and Karnataka, located in Southern India. The proposed model is compared with the state-of-the-art classification models in terms of various evaluation metrics for performance evaluation.