Harvest Guard: Crop Loss Detection Enhanced with Inception-Based Models

Diseases of plants are a major risk factor for world agriculture, causing severe crop loss and decreased food yields. This paper discusses the usage of deep learning models, that is InceptionV3 and InceptionResNetV2, in plant disease classification using the PlantVillage dataset. Preprocessing methods like the removal of duplicates and blur from images are performed to improve the performance of the model, and then data augmentation is used for enhanced learning. The data is systematically split into training, validation, and test sets to facilitate proper evaluation. Performance of the model is evaluated based on accuracy, precision, recall, and F1-score. The outcomes reveal that data augmentation significantly improves classification performance, with InceptionResNetV2 performing better than InceptionV3 in accuracy. Additionally, visual inspection of training patterns and mistakes gives insight into the strengths and weaknesses of the model. This study showcases the capability of deep learning in early detection of plant diseases and can lead to less crop loss and increased agricultural productivity.