Computer Vision System in Python

Comprehensive Computer Vision System and Useful Modules (vision-cnt) using OpenCV in Python. It includes:

• Mono Camera Model
• Stereo Camera Model
• Mono Calibration Module
• Stereo Calibration Module
• Image Processing Module
• Color Detection Module
• Target Box Regulator
• Tracking Module
• Transformation Module
• FPS Module
• Color Publisher

Author: Mehmet KAHRAMAN | Date: 12.03.2025 | Last Update: 28.04.2025

Requirements

Required packages can be installed with Pip. Ensure that pip is upgraded:

python3 -m pip install --upgrade pip

You can install all packages from requirements.txt:

• Numpy
• OpenCV
• Dlib

pip install -r requirements.txt

Installation

Upgrade pip version and install vision-cnt package with setup.py

pip install --upgrade setuptools wheel
pip install .

Calibration

You need to calibrate the camera first. Calibration images and npz data will be saved in calibration folder. Run Mono Camera Calibration example:

python3 examples/mono_calibration_example.py

Camera initializing...

Calibration initializing...

...

Color Detection & Tracking Example

You can add color targets and run color tracking example:

python3 examples/color_tracking_example.py

Also object tracking will work

Delta Color Publisher

The system enables that publish each unique color objects with their information to any specific topic using zeromq messages in order to make them accessible for delta robots.

{
    "id": 1,
    "color": 0,
    "position": [0.0, 0.1, 0.2],
    "velocity": [0.1, 0.0, 0.0]
}

Regulating Target Boxes

The algorithm can regulate close target boxes automatically.