An Enhanced and Lightweight YOLOv8-Based Model for Accurate Rice Pest Detection

The detection of rice pests plays a crucial role in enhancing agricultural productivity by enabling early interventions and reducing crop losses. This project presents an enhanced and lightweight pest detection system based on the YOLOv12 model, designed to accurately identify and classify pests in rice crops. The system utilizes YOLOv12, a state-of-the-art deep learning-based object detection model, to efficiently process high-resolution images and detect various pest species. The model is fine-tuned to balance accuracy with computational efficiency, making it suitable for deployment on resource-constrained devices in real-world agricultural settings. The project leverages transfer learning and data augmentation techniques to improve model robustness, ensuring reliable pest detection across different environmental conditions. Trained and evaluated on a custom dataset of rice pest images, the model's performance is assessed using metrics such as accuracy, precision, recall, and F1-score. The proposed system offers a scalable and efficient solution for pest detection, which can be integrated into smart farming systems to support timely pest management and contribute to sustainable agricultural practices.

Keywords: YOLOv12, Rice Pest Detection, Deep Learning, Object Detection, Image Processing, Transfer Learning, Agricultural Automation, Precision Agriculture, Pest Management, Sustainable Farming, Computer Vision, Model Evaluation.