Paddy insect identification using deep features with lion optimization algorithm.

Area of Science:

• Agricultural Science
• Computer Science
• Artificial Intelligence

Background:

• Pests cause substantial global rice yield losses, estimated at 20%.
• Early detection of paddy insects is crucial for mitigating these economic impacts.
• Existing insect identification systems lack integrated feature optimization with Deep Learning and Machine Learning.

Purpose of the Study:

• To develop a framework for prompt detection and categorization of paddy insects using advanced AI techniques.
• To enhance paddy insect image datasets through pre-processing and feature selection.
• To improve the accuracy and efficiency of paddy insect diagnosis in agricultural fields.

Main Methods:

• Gathering and categorizing a paddy insect image dataset.
• Applying pre-processing techniques like augmentation and image filtering.
• Utilizing 5 pre-trained Convolutional Neural Network models for feature extraction.
• Implementing feature selection methods: Principal Component Analysis (PCA), Recursive Feature Elimination (RFE), Linear Discriminant Analysis (LDA), and Lion Optimization.
• Employing 7 Machine Learning algorithms for insect identification.

Main Results:

• The proposed framework successfully detects and categorizes paddy insects from images.
• Feature vectors extracted using ResNet50 combined with Logistic Regression and PCA achieved the highest accuracy of 99.28%.
• The integration of feature optimization techniques significantly improved diagnostic capabilities.

Conclusions:

• The developed AI framework offers a highly accurate and efficient solution for paddy insect diagnosis.
• This approach has the potential to significantly reduce crop losses in paddy cultivation.
• The study highlights the importance of combining Deep Learning, Machine Learning, and feature optimization for agricultural pest management.