A deep learning model for predicting risks of crop pests and diseases from sequential environmental data.

Area of Science:

• Agricultural Science
• Machine Learning
• Environmental Monitoring

Background:

• Crop pests and diseases significantly reduce agricultural productivity.
• Early detection and prevention are crucial for effective crop management.
• Machine learning models are increasingly used for predicting crop issues using diverse data.

Purpose of the Study:

• To develop a deep learning model for predicting crop pest and disease risk.
• To utilize readily available environmental growth data for prediction.
• To demonstrate the model's applicability across various crops.

Main Methods:

• Utilized deep learning techniques to analyze historical environmental data.
• Included air temperature, relative humidity, dew point, and CO2 concentration as input features.
• Trained and validated the model on large-scale public datasets for multiple crops.

Main Results:

• Achieved high predictive performance with an average Area Under the Receiver Operating Characteristic Curve (AUROC) of 0.917.
• Successfully predicted the risk score for crop pests and diseases.
• Demonstrated the model's capability to identify potential disease outbreaks.

Conclusions:

• The proposed environmental data-based crop disease prediction model shows high accuracy.
• The model and learning framework are potentially universally applicable to various crops and facilities.
• This approach can significantly aid in proactive pest and disease prevention strategies.