Real-Time Tiger Intrusion Detection System Using Machine Learning and IoT

Tiger intrusion into human settlements poses serious threats to both wildlife conservation and human safety. These tigers usually venture into human regions resulting in loss of wild stock and human life. This study presents a real-time tiger intrusion detection system using IoT and machine learning algorithms for effective wildlife monitoring and conflict mitigation. This enables the people to safeguard themselves from the threat accordingly. This article aims to use computer vision models like mobileNet which can identify the intrusion of the tigers into an fixed region using CCTV cameras. The system integrates PIR motion sensors, thermal cameras, and acoustic sensors to detect tiger movements in real time. The collected data is transmitted via LoRaWAN and GSM-based IoT networks to a cloud-based platform for processing. A trained Convolutional Neural Network (CNN) and Random Forest classifier analyze the sensor data, achieving an accuracy of 94.6% in distinguishing tigers from other animals. Field tests in tiger-prone regions of India demonstrated a false positive rate of 3.8% and an average detection time of 2.3 s. Upon confirmation of tiger presence, real-time alerts are sent to forest officials and local communities for immediate action. The results highlight enhanced detection efficiency, reduced false alarms, and faster response times, ensuring improved conservation efforts and human safety.