CivicEye – Ubuntu (Docker Edition)

Features

• Real-Time Helmet Detection: Detects individuals as With Helmet (green bounding box) or Without Helmet (red bounding box) using a YOLOv8 model.

• Multiple Video Sources: Supports:

  • Local video files (.mp4, .avi, etc.)
  • Connected webcams
  • RTSP streams (CCTV or IP cameras)

• Frame Saving (‘Without Helmet’ detections):

  • Online Mode: Saves detection info and frames to the CivicEye cloud (civiceye.my) using your credentials.
  • Offline Mode: Saves detection frames locally under Detects/ inside the container or mounted volume.

• Adjustable Detection Speed: Choose between ‘Fast’, ‘Balanced’, and ‘Quality’ to manage performance vs. accuracy.

  • Fast (Skip=15) → Higher speed, more CPU.
  • Balanced (Skip=30) → Default balance.
  • Quality (Skip=60) → Fewer frames, lowest CPU.

• User Authentication:

  • Online Mode: Uses your civiceye.my login.
  • Offline Mode: Uses fixed local credentials (admin@civiceye / civiceye).

• Session Persistence: Online sessions remain active until logout.

• Modern GUI: Built using Tkinter and ttkbootstrap (supports Light/Dark themes).

• Containerized Setup: Runs fully inside Docker for isolation and easy deployment.

---

System Requirements

Component	Requirement
OS	Ubuntu 22.04+ or other Linux with X11 display
CPU	Intel i3 or higher
RAM	4GB minimum (8GB recommended)
Storage	~5GB free space
Docker	Latest version installed (sudo apt install docker.io)
GPU (optional)	NVIDIA GPU with CUDA if YOLO GPU acceleration is needed
Internet	Required for setup and Online Mode

---

Installation (Docker Setup)

1. Clone or Download the Project

git clone https://github.com/<your-org>/CivicEye.git
cd CivicEye

Ensure the directory contains:

• Dockerfile
• requirements.txt
• DATA/ folder with civiceye.py

---

2. Build the Docker Image

sudo docker build -t civiceye .

This might take a while , so grab a coffee

---

3. Run the Container

Run With GUI (local display)

If you are running on a Linux desktop:

xhost +local:root
sudo docker run -it --rm \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    civiceye

---

4. (Optional) GPU Support

If you have an NVIDIA GPU:

• Install the NVIDIA Container Toolkit
• Then run:

sudo docker run -it --rm \
    --gpus all \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    civiceye

---

Usage Guide

Login

1. Launch CivicEye (container starts GUI).

2. Choose:

   • Online Mode: Use your civiceye.my credentials.

   • Offline Mode:

     Email: admin@civiceye
     Password: civiceye
     

3. The app checks for internet connectivity before online login.

---

Main Interface

• Menu Button (☰): Opens side navigation.

• Theme Toggle: Switch between light (‘litera’) and dark (‘darkly’).

• Detection Speed: Choose from Fast, Balanced, Quality.

• Video Source:

  • Browse File: Select local video file.
  • Use PC Camera: Open default webcam.
  • Connect CCTV: Enter RTSP URL.

• Control Buttons:

  • Start Detection: Begin YOLO processing.
  • Stop Detection: Halt detection.

• Video Preview: Displays bounding boxes:

  • Green = With Helmet
  • Red = Without Helmet

• Status Bar: Shows current state messages.

---

Detection Outputs

• Online Mode: Saves detection data to the CivicEye server.

• Offline Mode: Saves JPEG images under:

  Detects/detection_YYYYMMDD_HHMMSS_ms_frame_XXX.jpg
  

  inside /app/Detects/ (or your mounted volume).

---

Technical Overview

Component	Description
Language	Python 3.10+
Core Libraries	Ultralytics YOLO, OpenCV, Tkinter, ttkbootstrap, Pillow, Requests
GUI Framework	Tkinter + ttkbootstrap
Containerization	Ubuntu 22.04 base, Python 3.10, Docker-managed
Optional GPU Support	CUDA runtime (via nvidia/cuda base image)

---

Uninstallation

To remove the Docker image:

sudo docker rmi civiceye

To remove all unused containers and images:

sudo docker system prune -a

---