Unmanned Aerial Surveillance and Tracking System in Forest Areas for Poachers and Wildlife

Poaching is morally wrong and has ecological consequences. Recent times have witnessed a large number of horrific poaching cases happening in India. Asiatic Lion, Bengal Tiger, Kashmir Red Stag, Blackbuck, etc., are a few species in India that have already been driven to endangerment. This paper proposes a UAS-based solution to help forest authorities prevent poaching activity. A wireless sensor network in which each node consists of a Peripheral Interface Controller (PIC) microcontroller, acoustic sensor, and Pyroelectric InfraRed (PIR) sensor is set up in wildlife hotspots for detecting intruders. The drone is deployed to the location of the triggered node. Using a robust two-fold collision avoidance system, the drone travels autonomously while avoiding obstacles in the dense forest cover. The UAS is programmed to detect poachers and animals using a custom-trained You Look Only Once (YOLO) model, and if poachers are detected, an alert is raised to the authorities. Simulations are done to validate navigation, collision avoidance, and poacher detection in the forest landscape. Proof of Concept of the proposed method was done on a hexacopter, which uses a Raspberry Pi 4B as the onboard companion PC for autonomous flight. The YOLO-based detection model’s ability to accurately identify items across various categories is reflected in its mean Average Precision (mAP) score of 0.914. In contrast, its F1 score of 0.88 highlights its ability to balance precision and recall. The prototype designed to illustrate the proposed algorithms was powered by a 2200 mAh battery, allowing a flight duration of around 7 minutes.