| dc.description.abstract | Detecting and classifying plant diseases in time and accurately is essential in food security to prevent
the spread and increase crop productivity. Traditional methods for detecting diseases require
knowledge in agriculture field and takes a lot of time for analyzing in laboratory which are not
efficient methods for farmers. In recent years, deep learning and especially CNN, shows its reliability
in detecting and classifying plant disease images which provides rapid diagnostic and does not rely
on expert knowledge. Therefore, this study focusses on developing a deep learning CNN model to
classify 9 diseases and 1 healthy tomato plant. Real time data augmentation is also applied to create
more variations and make the dataset more diverse which helps the model learn more features of the
same images and increase the model's generalization. The model is then compared with five other
pretrained models: Resnet50, VGG19, VGG16, InceptionV3, and MobilenetV2 to shows it
superiority performance with 99.05% accuracy, while the proposed model is the lightest one which
has the least number of parameters. | en_US |
