Paper Key : IRJ************628
Author: Mr. Kapalavayi Ramesh Babu,Rongala Ravindhra Naidu,Mandapalli Mounish,Yendluri Nithin,Shaik Mohammad Suhel
Date Published: 02 Apr 2025
Abstract
This study presents a multi-class classification approach for identifying plant leaf diseases using a combination of deep convolutional neural networks (CNN) and Local Binary Pattern (LBP) feature fusion. The method first classifies the plant species, such as Apple, Tomato, or Grape, and then classifies the specific disease affecting the plant, including Black Rot, Scab, Cedar Rust, and others. The process begins with an input image, which undergoes noise removal and restoration. LBP features are then extracted from the image, and the data is organized into a datastore, with labels split by proportion. Augmented image data is generated in batches to enhance training. A CNN is employed with various layers and training options to classify the plant species. After determining the plant species, the CNN undergoes additional training to classify the specific disease type. This dual-stage classification system improves accuracy by leveraging both deep learning and texture-based features, offering a robust solution for early and precise detection of plant diseases.
DOI Requested