3D Computer Vision (SoSe2023)

Some code examples used or demonstrated in the 3D Computer Vision course in summer term 2023

Prerequisites

First we need to download and install Python 3 from python.org. The recommended version is 3.10.7 or above. To inspect the currently installed version we can use the following terminal command:

python --version

Ensure that the correct version is used in the terminal as well as the selected interpreter.

---

Usually the dependency management tool pip is included in the Python installation. If it is not, download and install it from pip.pypa.io. Version 19.3 or above is recommended. We can again use a terminal command to check the version:

pip --version

Create an virtual environment in order to avoid installing the necessary modules globally. In the project root folder execute the following terminal command:

python -m venv .venv

The will create the hidden folder .venv in your current project folder. Next we need to enable the VENV. The VENV will only be active for the current terminal session. So closing the terminal and re-opening it, will disable the VENV.

Always make sure to enable it before you start developing.

# For Windows use
.venv\Scripts\Activate.ps1

# For Linux
source .venv/bin/activate

Special care is needed when using Windows. Please consult the following [guide][venv-guide] for more information.

---

On both Operation Systems the VENV can be deactivated with the following terminal command:

deactivate

Installing dependencies

Installing the required dependencies is straight forward:

pip install .

This will load the defined dependencies in pyproject.toml and install them inside the VENV.

Content

3DCV_Qt_Stereo.py is a tool to illustrate the epipolar geometry and shows how to compute the fundamental matrix. On start the fundamental matrix is calculated from features in two images. You can click points in one image and the corresponding epipolar line is drawn in the other.

The images can also be rectified.

In the lower part, the features including their matches can be shown as well as all epipolar lines for all found feature points.