Enhancing Betel Leaf Disease Detection Integrating Dcnn and Rpo Optimization for Accurate Classification

Abstract

Betel leaf, renowned for its cultural and medicinal importance, faces severe threats from various diseases, jeopardizing its cultivation and economic viability. Detecting diseases in betel leaves presents challenges hindering accurate diagnosis. This study addresses these challenges using a comprehensive approach. Initially, raw images are standardized to a pixel size and subjected to pre-processing using Contrast Limited Adaptive Histogram Equalization (CLAHE). Subsequently, segmentation techniques, including Gaussian Mixture Model (GMM) and Fuzzy C-Means (FCM), partition the images into meaningful segments based on intensity, color, and texture criteria, facilitating the extraction of relevant information and delineation of regions of interest. The segmented images undergo data augmentation to ensure balanced representation across classes and enhance model robustness through techniques like horizontal and vertical flipping. Feature extraction is performed using the Gray-Level Co-occurrence Matrix (GLCM) method on augmented images. Extracted features are then inputted into a classification phase, utilizing a Deep Convolutional Neural Network (DCNN) optimized by the Red Panda Optimization (RPO) algorithm. Through this methodology, the study seeks to improve disease detection accuracy in betel leaves, contributing to the preservation of their cultivation and economic significance. Experimental results demonstrate the effectiveness of the proposed approach in accurately classifying betel leaf diseases.