Integrated Forest Fire Detection System for Identifying Living Beings Using Drones by Employing Custom TrainedYOLOv5 Model

The substantial threat of forest fires impacts both the environment and humankind. Early detection of these fires plays apivotal role in effective fire fighting and minimizing damage. Addressing the challenges of identifying living beings amidst forest fires is crucial for effective emergency response. This research proposes a comprehensive approach by implementing Unmanned Aerial Vehicles (UAVs) integrated with advanced object detection technology to locate and identify animals and humans during such disasters. By employing a custom-trained Efficient YOLOv5 model specifically optimized for aerial views, the system achieves a mean Average Precision (mAP) of 0.828 at a confidence threshold of 0.5, demonstrating a reliable performance in detecting various living beings. Integration of this technology into an alert system enables real-time updates on the locations of living beings, facilitating timely and efficient rescue efforts. The proposed methodology not only enhances the effectiveness of emergency response teams but also contributes to minimizing the loss of life and property caused by forest fires.