No abstract

No abstract

The citrus industry depends on the early identification of fungal infections, since a few infected fruits may spread the disease to a whole batch, resulting in substantial economic losses. In recent years, deep learning has played a significant role in the automated identification and categorization of illnesses in vegetables and fruits. This has increased the quality and quantity of vegetables and fruits. Numerous illnesses have a negative influence on the quality of citrus crops. Different pre-trained CNN models were employed in this research to identify and classify citrus diseases. The different CNN Models are compared with five pre-trained CNN models for detecting citrus diseases. Different combinations of training and learning methods are used with pre-trained architectures, such as VGG16, InceptionNet, ResNet, NasNet, MobileNet, and CNN for disease detection. A dataset of around 1500 images of diseases and healthy citrus leaves has been collected from different sources. The simulation shows that, of all the models, the MobileNet architecture is the most accurate, with an accuracy rate of 96%.