Abstract :

Plant Diseases Significantly Affect Agricultural Productivity And Food Security By Reducing Both Crop Yield And Quality. Early Detection And Accurate Diagnosis Are Essential To Prevent Large-scale Crop Losses And Support Sustainable Farming Practices. With Rapid Advancements In Artificial Intelligence And Computer Vision, Automated Plant Disease Detection Systems Have Become Valuable Tools For Identifying Plant Diseases Using Leaf Images. These Intelligent Systems Help Farmers And Agricultural Experts Detect Diseases Quickly And Make Timely Crop Management Decisions. Traditional Plant Disease Detection Methods Mainly Rely On Manual Inspection By Agricultural Experts Or Laboratory-based Analysis. These Methods Are Time-consuming, Require Specialized Knowledge, And Are Not Practical For Large-scale Agricultural Monitoring. Conventional Machine Learning Approaches Have Attempted To Automate Disease Detection, But They Often Depend On Handcrafted Features And Limited Datasets. As A Result, They Struggle To Capture Complex Visual Patterns In Plant Leaf Images And May Produce Unreliable Results Under Varying Environmental Conditions. To Overcome These Limitations, This Research Proposes A Transformer-Driven Hybrid Pipeline For Scalable Plant Disease Diagnostics, Integrating Deep Learning, Machine Learning, And Explainable Artificial Intelligence (XAI). The System First Applies An XAI Module To Verify Whether The Uploaded Image Contains A Valid Plant Leaf And Extracts Contextual Attributes Such As Plant Type, Health Condition, And Dominant Color. After Validation, Deep Feature Extraction Is Performed Using DenseNet121, Which Captures Spatial Features, And A Vision Transformer (ViT) That Model’s Global Contextual Relationships. The Extracted DenseNet121 Features Are Processed Using Classifiers Including Perceptron, Nearest Centroid Classifier (NCC), And An Ensemble-based DenseNet121 With Ensemble Neighbor Model (ENM) That Combines K-Nearest Neighbor (KNN), And Radius Neighbors Classifier (RNC). Additionally, DeepPercepNet (DPN) Integrates ViT Feature Embeddings With A Perceptron Classifier To Enhance Prediction Performance. The System Detects Multiple Plant Diseases Across Crops Such As Apple, Cherry, Corn, Grape, Peach, Pepper, And Strawberry, Including Healthy Leaf Conditions. A Tkinter-based Interface Enables Interactive Image Uploads And Predictions, While Model Performance Is Evaluated Using Accuracy, Precision, Recall, F1-score, Confusion Matrix, And ROC Analysis

Published:

24-4-1-2026

Issue:

Vol. 26 No. 4-1 (2026)

Page Nos:

835-845

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite