This repository contains a unified and modular implementation of YOLOv5-based object detection models tailored for critical safety and surveillance applications. The merged codebase integrates multiple detection capabilities into a single, streamlined framework:

🔥 Fire Detection

💨 Smoke Detection

🔫 Weapon Detection

🧍‍♂️ Human Fall Detection

🤼 Physical Encounter Detection

The system is designed for real-time performance and can be easily deployed on edge devices. Each detection module has been trained and optimized to ensure high accuracy in dynamic environments such as industrial sites, public areas, and smart buildings.

🚀 Fire it up on your machine and let AI keep an eye on safety.

  git clone https://[email protected]/reedling/yolov5_deployments.git

Use your bitbucket security keys to continue.

cd yolov5_deployments/
pip install -r requirements.txt2

fall | smoke | car | gun | fire

    python SaveModelLink.py --model <model_name> --url <model_download_url>

    python DownloadModelWeights.py --model <model_name> --name <model_weights_name>

Run YOLO detection by giving the video and image source, alert type, and display options.

python inference.py --source 'rtsp://dl-team:dl@[email protected]:554/ch5/main/av_stream.h264' --alert Fire --conf-thres 0.3 --imgsz 640