Code used to generate the results in the paper: Kimberly McGuire, Guido de Croon, Karl Tuyls and Hilbert Kappen, "Efficient Optical flow and Stereo Vision for Velocity Estimation and Obstacle Avoidance on an Autonomous Pocket Drone", Submitted to Robotics and Automation Letters (2016)
The code is developed for Matlab 2016a (the newest computervision library). It requires the Package 'Statistics and Machine Learning Toolbox' to output error box plots.
To start up the code, run Edgeflow_generate_results.m.
Before you start: The images used by these scripts are available at the github repository https://github.com/knmcguire/stereoboard_testing.git. This repository is linked as a submodule and can be downloaded by simply updated the submodules. For linux, type the following lines into the terminal when in the main Edgeflow_matlab folder:
git submodule init
git submodule update
Also this repo relies on the matlabpackes matlab2tikzz (https://www.mathworks.com/matlabcentral/fileexchange/22022-matlab2tikz-matlab2tikz)