Deep Learning based Real-Time Animal Detection using MobileNetV2

Abstract

In this modern world, Artificial intelligence has emerged as a powerful tool in wildlife conservation, agricultural land protection, and environmental monitoring. The study proposes a real-time animal detection and classification system by utilizing the MobileNetV2 deep learning model. MobileNetV2 is selected for its lightweight architecture, high computational efficiency, and superior accuracy, making it suitable for real-time applications. The model achieves an outstanding accuracy of 99.62%, which outperforms with other deep learning models. The proposed system enhances real-time monitoring by detecting and classifying animals in various environments, aiding in wildlife protection, mitigating human-animal conflicts, and securing agricultural lands. By providing instant alerts, the system enables rapid decision-making during emergency situations, ensuring safety and effective intervention. The system is designed to be scalable, cost-effective, and adaptable to different terrains, making it an ideal solution for large-scale deployment in conservation areas and farmlands.