Affiliation:
1. Department of Electronics Engineering (VLSI Design and Technology), Velammal College of Engineering and Technology (Autonomous), Madurai, India
Abstract
Sugarcane, vital to industries such as food, biofuels, and pharmaceuticals, is threatened by diseases like red rot, rust, yellowing, and mosaic, which require significant resources for detection and monitoring. This book chapter presents an innovative solution using DenseNet-121, a convolutional neural network to identification and classification of these leaf diseases. The study utilized a plant village dataset of 9,050 leaf images, including healthy and diseased samples, which were carefully organized and pre-processed to ensure uniformity. The enhanced DenseNet-121 model, trained on this dataset, demonstrated outstanding performance, with a precision of 0.96, recall of 0.977, F1-score of 0.98, and accuracy of 0.963. While effective, challenges like resource needs may limit implementation in constrained environments, yet it marks progress toward agricultural sustainability and food security.