Deep Learning Meets Agriculture: A Faster RCNN Based Approach to pepper leaf blight disease Detection and Multi-Classification

Abstract

Pepper Leaf Blight Disease (PLBD) is a widespread plant ailment that has a severe impact on pepper cultivation across the globe. The rapid detection and precise classification of PLBD severity levels are crucial for efficient disease control and optimal agricultural productivity. The present study introduces a novel model based on Faster region-based convolutional neural network (R-CNN) for the efficient detection and multi-classification of PLBD in pepper leaves. The dataset used for training and testing the model consisted of 10,000 images. The model’s performance was evaluated based on its detection accuracy and multi-classification accuracy, which were found to be 99.39% and 98.38%, respectively. The model’s computational efficiency was assessed and determined to be sufficient for deployment in real-time disease detection applications. The model’s average inference time of 0.23 seconds per image renders it appropriate for deployment in high-throughput disease detection applications. The study’s findings indicate that the faster RCNN model is a successful method for detecting and classifying PLBD in pepper leaves. This has the potential to enhance disease management and crop yield in pepper farming.