This repository contains the FAITH algorithm implementation on ROS.
Supporting paper: https://arxiv.org/abs/2102.12823
We provide the Matlab implementation for our FAITH method, as well as the 3 other methods that we used for comparison in our paper. Namely:
- The NESW method: Huang, R., & Ericson, S. (2018, June). An Efficient Way to Estimate the Focus of Expansion. In 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC) (pp. 691-695). IEEE.
- The Vec. Intersections method: Buczko, M., & Willert, V. (2017, June). Monocular outlier detection for visual odometry. In 2017 IEEE Intelligent Vehicles Symposium (IV) (pp. 739-745). IEEE.
- The Half-planes method: Clady, X., Clercq, C., Ieng, S. H., Houseini, F., Randazzo, M., Natale, L., ... & Benosman, R. B. (2014). Asynchronous visual event-based time-to-contact. Frontiers in neuroscience, 8, 9.
Each function takes as input the (x,y) coordinates of the PoIs that provided the flow vectors (u,v). The output is the FoE candidate (x_foe,y_foe) coordinates.
These ROS packages are meant for use with ROS 1 (Kinetic, Melodic). They require you to use the DVS240 event-based camera, along with the ROS drivers ().
- The
dvs_of
package is meant to output derotated flow from the DVS. - The
foe_estimator
package uses the optic flow provided by thedvs_of
package to estimate the FOE position (FAITH). - The
object_detection
package uses the FOE estimation to make a drone equipped with the DVS camera avoid obstacles.
For further information, please contact Julien Dupeyroux: j.j.g.dupeyroux@tudelft.nl