A novel hybrid inception-xception convolutional neural network for efficient plant disease classification and...

Area of Science:

• Agricultural Science
• Computer Science
• Artificial Intelligence

Background:

• Plant diseases and pests pose significant threats to global food security, leading to reduced crop yields and economic losses.
• Traditional laboratory diagnosis methods for plant diseases are often costly, time-consuming, and labor-intensive.
• Early and accurate detection of plant diseases is essential for effective management and mitigation strategies.

Purpose of the Study:

• To develop and evaluate a novel hybrid Inception-Xception (IX) Convolutional Neural Network (CNN) model for automated plant disease detection and classification (PDDC).
• To create a real-time AI application for plant disease identification using improved CNN, machine learning (ML), and computer vision techniques.
• To assess the performance of the proposed IX-CNN model against various datasets and classifiers.

Main Methods:

• A hybrid Inception-Xception (IX) model was designed, integrating inception and depth-separable convolution layers for multi-scale feature extraction and reduced complexity.
• The IX-CNN model was trained and validated using six diverse datasets: PlantVillage, Turkey Disease, Plant Doc, Rice Disease, RoCole, and NLB.
• Performance was evaluated using classifiers such as Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF).

Main Results:

• The IX-CNN model achieved high accuracy across multiple datasets, with Plant Doc, PlantVillage, and Turkey Disease datasets reaching 100% accuracy.
• The Rice Disease, RoCole, and NLB datasets demonstrated excellent performance with accuracies of 99.79%, 99.95%, and 98.64%, respectively.
• The developed AI application provides real-time, automated plant disease identification and classification.

Conclusions:

• The proposed hybrid IX-CNN model offers a robust and efficient solution for early plant disease detection and classification (PDDC).
• The model's high accuracy and real-time application capabilities support smart farming initiatives and contribute to mitigating threats to global food security.
• This AI-driven approach can significantly reduce reliance on traditional diagnostic methods, enabling faster interventions and improved crop management.