Published Paper Details:

Animal intrusion refers to the unauthorized presence of animals in areas such as farms, forests, or urban zones, often leading to damage, safety risks, or property loss. These incidents, frequently caused by wild or domestic animals, necessitate rapid and reliable monitoring systems. The integration of Artificial Intelligence (AI) has greatly enhanced these systems by improving accuracy, speed, and automation in surveillance. Machine Learning (ML) and Computer Vision (CV) approaches enable automatic detection and monitoring of animal movements via camera feeds or sensor data, assisting in wildlife protection and security measures. This study explores a deep learning-based approach to detect animal intrusions in diverse outdoor environments. Models like YOLOv5 and EfficientDet, trained using PyTorch and fine-tuned for real-time performance, are evaluated using key metrics such as 2 accuracy, precision, recall, F1-score, and mean Average Precision (mAP). Among the tested models, YOLOv5 achieved the highest accuracy, with a mean mAP of 75.7%, indicating strong real-time detection capabilities. The findings confirm that object detection models based 1 on deep Convolutional Neural Networks (CNNs) are highly effective in identifying both wild and domestic animals, supporting better wildlife monitoring and security systems.