Design of Smart System for Mitigating Wild Animal Intrusion in Agricultural Farms Using IoT and Deep Learning
A. Prasanth, M. S. Arunkumar, B. Senthil Kumaran, S. Deepa
Abstract
The recent development of the Internet of Things (IoT) and Deep Learning models have been enhanced to resolve day-to-day constraints in the agricultural farm. The main constraints are crop damage, human injury, property loss, etc. In the past, traditional methods such as human surveillance, dog bodyguards, and electric fences were utilized to safeguard the agricultural field. However, these methods provide only temporary solutions that fail to save wildlife and humans. Some mitigation techniques are needed to solve these problems. In this work, the wild animal intrusion detection system has been introduced to enhance the safety of wild animals and humans. A novel Adaptive Dot Dis-tance-based YOLOv9 (ADDY) model is employed in the proposed system to detect wild animal intrusions in agricultural farms. In the proposed ADDY model, the Dot Distance (DD) mechanism has been integrated with the YOLOv9 to enhance the decision-making process. The preliminary results show that the proposed ADDY model has a superior accuracy of 98.75% when compared to the existing detection techniques.