A Preliminary Study on YOLOv8-based Animal Detection for Farmland Protection

Wild animals entering agricultural lands frequently result in crop damage and pose safety hazards for farmers, particularly in areas adjacent to forests and wildlife pathways.Conventional protective strategies such as fencing, firecrackers, and manual surveillance often prove to be ineffective, costly, or unsafe.Recent advancements in deep learning and computer vision have made it possible to detect animals in real-time using camera systems.Among the available models, YOLOv8 strikes an effective balance between accuracy, processing speed, and ease of implementation, even on devices with limited computational capabilities.This paper introduces a preliminary study on a YOLOv8-based detection framework, initially aimed at simulated evaluation and later for potential use in agricultural settings.The goal is to explore a practical solution for early warning to minimize crop losses and foster safer coexistence between farming communities and local wildlife.