DroneAI

DroneAI is a local web-based GUI for labeling drone flight videos. It helps users upload a video dataset, label flight events, track progress, create training data, and review crash results through a browser-based interface.

This first version is focused on the local application workflow.

---

Requirements

Before running DroneAI, make sure you have:

• Python 3.10 or newer
• Git
• pip
• FFmpeg installed and added to PATH
• A virtual environment
• An Excel dataset file

Your Excel file should include these columns:

Persona Name
Youtube Link

---

Setup

Clone the repository:

git clone https://github.com/YOUR_USERNAME/DroneAI.git
cd DroneAI

Create a virtual environment:

python -m venv venv

Activate the virtual environment.

On Windows:

venv\Scripts\activate

On Mac/Linux:

source venv/bin/activate

Install dependencies:

pip install flask pandas openpyxl opencv-python numpy yt-dlp paho-mqtt

If the project includes a requirements.txt file, you can also run:

pip install -r requirements.txt

---

Run the Application

Start the Flask app:

python LabelGUI/app.py

Then open this address in your browser:

http://localhost:5000

---

Login

Use any username.

Default password:

droneai2025

Choose a role:

• Leader — can upload and manage datasets
• Team Member — can label videos from the queue (If you want to label videos, just log in as team member)

---

Basic Local Workflow

1. Log in as a leader.
2. Open Datasets.
3. Upload the Excel dataset.
4. Set the dataset as active.
5. Open the queue.
6. Start a video row.
7. Mark events while the video plays.
8. Return to the queue after the session finishes.
9. Confirm the row is marked as labeled.

---

Main Pages

• Dashboard — shows the current dataset and labeling progress
• Datasets — upload and manage Excel datasets
• Shared Queue — select videos to label
• Training — create labeled training data
• Crash Analysis — review crash verification results
• DB Tools — export or import the local database

---

Local Database

Progress is saved in the local SQLite database:

db/droneai.sqlite

If you want another machine to continue from the same state, export the database from DB Tools and import it on the other machine.

---

Author

Developed by Haider Baig.