crazyKhoreia

Brief description.

crazyKhoreia is a robotic perception system intended for UAVs (aka drones), which takes a digital image and converts it into a waypoint matrix using X, Y and Z coordinates in meters for UAVs choreography design.

A preprint paper of this project is available here.

If this work is useful to you, please cite us:

@article{Restrepo2023, author = "Santiago Restrepo and Victor Romero Cano", title = "{CrazyKhoreia, a robotic perception system for UAV path planning from digital images}", year = "2023", month = "1", url = "https://www.techrxiv.org/articles/preprint/CrazyKhoreia_a_robotic_perception_system_for_UAV_path_planning_from_digital_images/21836868", doi = "10.36227/techrxiv.21836868.v1" } 

How does it work?

crazyKhoreia takes a digital image and uses a vectorizing technic based on openCV findContours algorithm to obtain contours, then, it gets waypoints from those contours, after that, depending on the usage mode (either lightPainting or multi-UAV formation), it optimizes the waypoints and finds a feasible 2D path for travelling through all of them, or it computes safe trajectories for multi-UAV formation.

Light painting mode (class lightPainting)

Light painting is a photography technic consisting on capturing a long exposure photograph using a moving light source, the light source's movement creates an effect on the camera sensor which captures its trail, therefore, it "paints" all over the camera's frame, creating a painting optical effect.

Source: Rafael Rodrigo Perurena

Multi UAV formation mode (class multiDroneFormation)

A multi UAV formation is made by several UAVs grouped together in patterns to construct a figure visible with a naked eye. It's been used by companies such as Intel or DSS to create terrific light shows performed by UAVs like the Super Bowl's half-time shows. Source: Intel

Installation instructions.

Prerequisites.

Before using crazyKhoreia, please install the following dependencies.

$ pip3 install opencv-python
$ pip3 install numpy
$ pip3 install matplotlib
$ pip3 install scipy
$ pip3 install scikit-learn

Installing from Pypi with Python 3.

To install crazyKhoreia you'll need Python 3 and pip:

pip3 install crazyKhoreia

Installing from source.

If you'd like to develop, edit or have full access into the crazyKhoreia system, then you may install it from source.

git clone https://github.com/santiagorg2401/crazyKhoreia.git

Usage.

crazyKhoreia's usage it's pretty straightforward, you'll only need a digital image and a few parameters:

Take into account that lightPainting and multiDroneFormation classes creates an instance of the crazyKhoreia class in its constructor method.

To run the program, create an instance of the lightPainting class.

from crazyKhoreia.lightPainting import lightPainting

lp = lightPainting(dims, in_path, out_path, detail=0.05, speed=1.0, sleepTime=1.5, video=False, led=False)

Or if you want to execute multiDroneFormation.

from crazyKhoreia.multiDroneFormation import multiDroneFormation

mdf = multiDroneFormation(dims, boxShape, in_path, out_path, nmbr_drones)

After its execution you'll notice the output files within the set output path.

Trouble?

Start a new discussion if you have any question related to the project, but, if you have a technical issue or a bug to report, then please create an issue.

Authors.

• Santiago Restrepo García.

Tutor.

• Víctor Adolfo Romero Cano.