Paddy Leaf Disease Detection using Deep Learning

Abstract

Paddy leaf disease detection using deep learning refers to training the neural networks on images of paddy leaves to precisely identify diseases. It supports early detection, wrapped in an overall reduction of crop loss, providing windfall to agricultural productivity. The consequences of plant diseases are a significant limitation to agricultural productivity, and monitoring manually is usually cumbersome, unreliable, and time-consuming. The model ORB-DL is used to extract the key features for identifying plant diseases. Combined with advanced DL models, MobileNetV2, ResNet50 these features increase the accuracy and robustness of disease detection. Thermal Imaging is capable of detecting small alterations even before visible symptoms manifest and allows for event-driven management. Grad-CAM visualization techniques provide interpretability results that afford insight into model predictions and build up confidence in automated solutions. Our experiment will show that the combination of ORB-DL with these DL architectures outperforms existing methods while still providing superior accuracy and reliability. The objectives of this study are to employ some of the means of Artificial Intelligence, Deep Learning (DL), and Thermal Imaging in early disease detection and mitigation of shortcomings.