Requirements¶

OTVision is designed to run on most modern operating systems (Windows, Linux and Mac). However, a few prerequisites are required before installing OTVision.

Hardware prerequisites¶

OTVision runs on modern desktops and laptops (e.g. Intel i5+ of the last few generations, AMD Zen chips or Apple Silicon processors and 8 GB RAM).

If you want to detect road users in videos with OTVision on a regular basis, we strongly recommend purchasing a powerful desktop computer or workstation (>= 8+ cores, >= 16 GB RAM) with a Nvidia graphics card (>= GeForce 10XX Series, better is usually faster). Make sure that the Nvidia drivers as well as the NVidia CUDA Toolkit are installed and up-to-date to get the best performance.

Apple Neural Engine

Accelerated detection using the Apple Neural Engine is now supported. Detection on Apple silicon chips can utilize the Neural Engine for improved performance.

Software prerequisites¶

uv¶

OTVision uses uv as a fast Python package installer and resolver to manage dependencies. The installation scripts automatically use uv to install Python packages from the Python Package Index.

Additionally, uv can manage Python interpreters itself (download, install, and select specific versions), so pre-installing Python is not required. uv will get and use an appropriate Python interpreter automatically through the required version defined in the pyproject.toml file.

Please refer to the uv documentation for further information on how to install uv on your system.

ffmpeg¶

If you want to use videos recorded by OTCamera with OTVision, you need to convert the videos to .mp4 files first. For the conversion, we use the Open Source software ffmpeg.

Before using the convert.py script, make sure that ffmpeg is installed and available on the whole system:

WindowsLinuxmacOS

To install ffmpeg on your Windows system, please perform the following steps:

Download the file ffmpeg-git-full.7z from gyan.dev.
Unzip this file by using any file archiver such as 7zip in a folder of your choice (e.g., C:\ffmpeg).
Now, open a Command Prompt with administrator privileges and set the environment path variable for ffmpeg:
```
setx /m PATH "path_to_your_ffmpeg_folder\bin;%PATH%"
```
where path_to_your_ffmpeg_folder represents the folder that you have ffmpeg unzipped in.

If you unzipped to C:\ffmpeg, for example:
```
setx /m PATH "C:\ffmpeg\bin;%PATH%"
```
Restart your computer and verify the installation by running
```
ffmpeg -version
```

Ubuntu repositories¶

If you use Ubuntu, you can install ffmpeg using the official Ubuntu repositories.

sudo apt install ffmpeg

Manual installation¶

For the manual installation of ffmpeg on Linux or if you use another distribution, please refer to the instructions on the developers website.

Homebrew¶

For the installation using Homebrew, use the following command in the terminal:

brew install ffmpeg

Manual install¶

For the manual installation of ffmpeg on macOS, please refer to the instructions on the developers website.

Nvidia CUDA (optional)¶

If you intend to use OTVision on a Windows or Linux PC with a modern Nvidia graphics card, download and install version 12.8 of the NVIDIA Cuda Toolkit.