Tomato Diseases and Pests Detection Based on Improved Yolo V3 Convolutional Neural Network

Area of Science:

• Agricultural Science
• Computer Vision
• Artificial Intelligence

Background:

• Tomato cultivation faces significant yield losses due to diseases and pests.
• Timely and accurate identification of these threats is crucial for effective management.
• Traditional methods for disease and pest recognition are often labor-intensive and less accurate.

Purpose of the Study:

• To develop an efficient and accurate deep learning-based system for identifying tomato diseases and pests.
• To improve upon existing object detection models for real-time application in agricultural settings.
• To provide a practical solution for farmers to mitigate crop damage and increase tomato yield.

Main Methods:

• Utilized deep learning object detection, specifically optimizing the YOLOv3 model.
• Developed a dataset of tomato diseases and pests captured in natural environments.
• Implemented an image pyramid approach to enhance multi-scale feature detection within the YOLOv3 model.

Main Results:

• Achieved improved detection accuracy and speed for tomato diseases and pests.
• Successfully enabled accurate localization and categorization of threats in real-time.
• Demonstrated the effectiveness of the optimized YOLOv3 model in natural environmental conditions.

Conclusions:

• The optimized YOLOv3 model provides a breakthrough in real-time tomato pest and disease recognition.
• This technology offers significant potential for intelligent recognition and engineering applications in agriculture.
• The study contributes valuable insights for reducing crop losses and enhancing tomato production.