DeepPlantGuard: An Intelligent Framework for Plant Disease Detection in Smart Farming Systems

Abstract

Leaf diseases are a major contributor to food and agricultural insecurity globally. The primary means of managing leaf disease through successful detection and classification of leaf diseases from field images. In this paper, we present a framework to automatically detect and classify plant leaf diseases, using a DCNN model. We further employ the principles of transferlearning; by fine-tuning previously trained deep learning models (e.g., VGG16, ResNet50, InceptionV3) with a total of 38 different types of diseases across 14 from different types of plants in our dataset. To support feature extraction from field images affected by variations in illumination and background interference, we utilize advanced image processing workflows (e.g., noise filtering, segmentation and image augmentation).Experimental evaluation achieved classification accuracies of 98.42% on the PlantVillage dataset and 96.78% on the field dataset, demonstrating the effectiveness of the proposed framework under both controlled and real-world conditions. The importance of each element in our model has been evaluated via ablation experiments. Our model will make automatic plant disease detection possible on mobile devices in the future.