A Deep-Learning Model for Real-Time Red Palm Weevil Detection and Localization.

Area of Science:

• Agricultural Entomology
• Computer Science
• Machine Learning

Background:

• Red Palm Weevils (RPW) are a major global threat to palm trees, particularly in the Middle East.
• Early identification of RPW is vital for date production and preventing widespread infestation.

Purpose of the Study:

• To develop a deep-learning model for accurate identification and classification of RPW.
• To discriminate RPW from other insects in palm tree habitats.

Main Methods:

• Utilized a region-based convolutional neural network (R-CNN) for RPW detection and localization.
• Employed a convolutional neural network (CNN) for feature extraction within bounding boxes.
• Integrated classification and regression layers for precise RPW identification.

Main Results:

• Achieved 100% accuracy in early-stage detection of RPW in infested palm trees.
• The model enables rapid and precise identification of the pest.

Conclusions:

• The developed deep-learning model offers an efficient, low-cost solution for RPW detection and classification.
• This technology can be deployed in large-scale agricultural settings, like the Al-Qassim region, to protect palm trees.