Hybrid Neural Network Approach for Tea Leaf Disease Detection Using Pelican and Mayfly Optimization Algorithms

Abstract

This study addresses the problem of plant diseases and the difficulty of detecting them, and it presents a unique technique for the automatic detection of tea leaf diseases by combining neural networks and optimization techniques. Our research uses a curated database of tea plant leaf photographs that includes healthy and diseased specimens. The neural network (CNN) is trained and fine-tuned using optimization algorithms. To increase disease identification accuracy, we used a hybrid novel optimization algorithm called (POA-MA) which is Pelican Optimization Algorithm (POA), and Mayfly Optimization Algorithm (MA) for feature selection, followed by classification with Support Vector Machine (SVM). The suggested mechanism performance is evaluated using accuracy, MSE, F-score, recall, and sensitivity measures. The suggested CNN-POAMA hybrid model yielded 94.5%, 0.035, 0.91, 0.93, and 0.92, respectively. This study advances precision agriculture by establishing a strong framework for automated detection, allowing for early intervention, and eventually enhancing tea crop health.

Keywords

• support vector machine (svm)
• convolutional neural network (cnn)
• visual geometry group (vgg-16)
• pelican optimization algorithm (poa)
• mayfly optimization algorithm (ma)