Smart Boundary Monitoring System for Wildlife Containment and Human Safety

This research intends to improve wildlife management and monitoring by utilising deep learning for effective animal detection in a variety of environments, including farming zones, reserves, and rainforests. Conventional techniques, such manual observation and simple image processing, are frequently time-consuming, labor-intensive, and prone to errors. The yolov5 object identification method, which is well-known for its accuracy and speed of real-time detection, is used in this research to discuss these limits. The system consists of a python-based back end with flask and a front end constructed with html, css, bootstrap, and javascript. Yolov5 uses species labels and bounding boxes to identify and categorise animals in video streams. To guarantee accuracy, the model is trained on a variety of animal datasets. Transfer learning is used to enhance performance with sparse data. Additionally, realtime SMS and sound alerts notify personnel upon detection, supporting rapid response, conservation, and the prevention of human-wildlife conflict.