Rice leaf disease classification using a fusion vision approach.

Area of Science:

• Agricultural Science
• Computer Vision
• Plant Pathology

Background:

• Rice is a global staple crucial for food security.
• Plant diseases pose a significant threat to rice yield and quality.
• Early disease detection is vital for effective crop management.

Purpose of the Study:

• To develop a novel and accurate method for early rice disease detection.
• To integrate deep learning for feature extraction and machine learning for classification.
• To evaluate the performance of the proposed model against existing classifiers.

Main Methods:

• Utilized the Fusion Vision Boosted Classifier (FVBC) approach.
• Integrated VGG19 for image feature extraction and LightGBM for classification.
• Trained and validated the model on a dataset of 2627 rice leaf images.

Main Results:

• Achieved high accuracy rates: 97.78% on training, 97.5% on validation, and 97.6% on the test set.
• Demonstrated superior performance compared to other classifiers like Softmax.
• Optimized model performance through hyperparameter tuning (learning rate, tree depth).

Conclusions:

• The FVBC model provides an effective, non-invasive solution for early rice disease detection.
• The model's scalability supports widespread application in agriculture.
• Timely disease identification empowers farmers to enhance crop productivity and quality.