Vanrakshak AI: AI Driven Wildlife Sanctuary Monitoring System

Abstract

Monitoring wildlife in forest areas is difficult and requires continuous observation. Traditional monitoring methods depend on manual patrol or basic camera systems which cannot detect animals automatically. This project proposes a smart monitoring system called Vanrakshak AI that detects animals using computer vision techniques. The system captures video from surveillance cameras and processes each frame to identify animal movement. Motion detection is used to reduce unnecessary processing and improve efficiency. The YOLO object detection model is used to detect animals, while DeepSORT tracking helps track animals across frames. When an animal is detected, the system stores the image, updates the database, and sends alerts to the monitoring dashboard. In some cases, notifications can also be sent through Telegram to inform forest authorities. The system helps in real-time monitoring and can assist forest departments in protecting wildlife and preventing human-animal conflicts.

Introduction

It explains that traditional wildlife monitoring methods rely heavily on manual observation and basic camera systems, which are slow, labor-intensive, and unable to provide real-time alerts, especially in large forest areas or low-light conditions. This delay often increases the risk of dangerous encounters between humans and wildlife.

To solve this problem, the proposed system uses computer vision and deep learning techniques to automatically detect and track animals from live camera feeds. It combines motion detection, YOLO-based object detection, and DeepSORT tracking to identify animals, follow their movement across frames, and store detection data with timestamps.

The system architecture includes several modules:

• Input acquisition from forest surveillance cameras
• Frame processing with motion detection for efficiency
• Animal detection using YOLO
• Tracking using DeepSORT to maintain identity across frames
• Database storage of detected images and logs
• Notification system that sends real-time alerts via dashboard and Telegram

The flow of the system begins with capturing video frames, detecting motion, running AI-based detection when needed, tracking animals, storing results, and triggering alerts when necessary.

The system also includes a web-based dashboard where users can monitor camera feeds, view alerts, generate reports, and manage camera settings and user access.

Conclusion

The proposed work aims to improve forest monitoring, which is generally slow and depends a lot on manual observation. In real situations, it is difficult for forest guards to continuously keep track of animal movement, and some important activities may get missed. To solve this problem, the proposed system uses YOLOv11 along with DeepSORT tracking to detect and follow animals automatically. It also sends alerts through Telegram and provides a live dashboard, helping authorities respond faster when required. This reduces manual effort and makes monitoring more efficient. In future, the system can be further improved to run on low-power devices so that it can be used in remote forest areas. Overall, the proposed work provides a simple and practical solution for real-time wildlife monitoring.

References

[1] L. Chen, “YOLO-SAG: An Improved Wildlife Object Detection Algorithm Based on YOLOv8,” Image and Vision Computing, Elsevier, 2024. [2] Z. Ma, Y. Li, and H. Wang, “Wildlife Real-Time Detection in Complex Forest Scenes Based on Lightweight YOLO Network,” Remote Sensing, vol. 16, no. 8, 2024. [3] Y. Zhu, J. Liu, and X. Zhang, “YOLO-WildASM: An Object Detection Algorithm for Protected Wildlife Monitoring,” Animals, vol. 15, 2025. [4] N. Nagaraj, “Automated Approach for Wildlife Detection and Tracking Using YOLOv8 Deep Learning Algorithm,” Springer Nature Computer Science, 2025. [5] D. Velasco-Montero, M. López-Nores, and J. Pazos-Arias, “Reliable Integration of Artificial Intelligence into Camera Trap Systems for Wildlife Monitoring,” Ecological Informatics, Elsevier, 2024. [6] F. Schindler, S. Steinhage, and K. Brunn, “Action Detection for Wildlife Monitoring Using Camera Trap Videos,” Applied Sciences, vol. 14, 2024. [7] R. A. Rajagukguk, A. Nugroho, and B. Siregar, “Deep Learning for Visual Animal Monitoring: Detection, Tracking and Behavior Analysis,” Smart Agricultural Technology, 2025. [8] T. Wang, Y. Sun, and L. Zhao, “Nighttime Wildlife Object Detection Based on YOLOv8-Night Model,” IET Electronics Letters, 2024.

Copyright

Copyright © 2026 Aditya Pachchigar, Atharva Yengatwar, Ansh Malviya, Aman Dambhare, Atharva Sagne, Dr. Ruchi Kadwane. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.