Installation

Prerequisites

DroneVis requires python 3.7+, PyTorch >= 1.13.0, TorchVision >= 0.14.0, and mxnet >= 1.9.0

Windows 10

Currently, the library is neighter tested nor built on Windows 10 platform. There are two options available though:

Use WSL to install Linux on Windows.
Fetch our docker image from DockerHub, and run it.
Create a Docker container with ubuntu >= 20.04, and install our library.

You can visit the section docker for more details.

Linux

The library is only heavily tested on Ubuntu >= 20.04.

Creating a virtual env

On ubuntu 20.04 and python 3.8:

Note

It is recommended to use a virtual environment to avoid packages conflict, but it is ok to omit the virtual env part.

$ sudo apt-get update # update current packages
$ sudo apt-get install -y virtualenv python3-virtualenv # install python virtualenv
$ virtualenv ~/dronevisvenv --python=python3.8 # create dronevis virtual env
$ source ~/dronevis/venv/bin/activate # activate virtualenv

Install using pip

To use DroneVis, you need to install it using pip:

(.dronevisvenv) $ pip install dronevis # install dronevis library

Development version

To contribute to DroneVis, with support for running tests and building the documentation.

$ git clone https://github.com/ahmedheakl/drone-vis
$ cd drone-vis
$ pip install -e .[docs]

Docker

You can now install docker on your platform, and pull our image from docker hub and start working right-away.

To pull our docker image ahmedheakl/dronevis:

$ docker pull ahmedheakl/dronevis

To run the image with interactive terminal:

$ docker run -it ahmedheakl/dronevis

You can now start running your favourites computer vision alogrithms.