Skip to content

Implentation of the FA-Harris algorithm, described in Ruoxiang Li et al. "FA-Harris: A Fast and Asynchronous Corner Detector for Event Cameras", IROS 2019.

Notifications You must be signed in to change notification settings

ruoxianglee/fa_harris

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

FA-Harris: A Fast and Asynchronous Corner Detector for Event Cameras

This code is the reference implementation described in the paper "FA-Harris: A Fast and Asynchronous Corner Detector for Event Cameras". This work was developed by Ruoxiang Li at National University of Defense Technology.

In this work, we present our event-corner detection method operating directly on the asychronous event-streams. The asychronous event corner detector, called FA-Harris, works based on the Surface of Active Events structure. We propose a global SAE construction and updating method and an efficient candidate selection and refinement strategy. And the FA-Harris algorithm runs 8 times faster than the novel event-based Harris detector. When considering the accuracy, our method achieves good performance compared to the previous methods.

This code has been tested with ROS kinetic on Ubuntu 16.04.

Video

FA-Harris detector

Publication

If you use this work, please cite the following publication:

Ruoxiang, Li and Dianxi, Shi and Yongjun, Zhang and Kaiyue, Li and Ruihao, Li. "FA-Harris: A Fast and Asynchronous Corner Detector for Event Cameras." IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019.

@inproceedings{li2019fa,
author = {Li, Ruoxiang and Shi, Dianxi and Zhang, Yongjun and Li, Kaiyue and Li, Ruihao},
year = {2019},
month = {09},
pages = {6223--6229},
title = {FA-Harris: A Fast and Asynchronous Corner Detector for Event Cameras}
}

Installation

Requirements:

  • Eigen 3 sudo apt-get install libeigen3-dev

ROS-related requirements:

Run

Clone the repository to your ROS workspace and build it:

$ cd /path/to/catkin_ws/src
$ git clone https://github.com/ruoxianglee/fa_harris
$ roscd fa_harris
$ catkin build --this

Connect your DAVIS camera and launch the following file:

$ roslaunch fa_harris corner.launch

Or, you can also play a rosbag file from the Event Camera Dataset:

$ roslaunch fa_harris corner.launch rosbag_flag:=1 rosbag_path:=/path/to/ros_bag.bag

Contact

Please, create an issue if you have questions or bug reports. If you come up with any improvements, please create a pull request. Alternatively, you can also contact me at ruoxianglee@163.com.

Acknowledgements

Many thanks to Ignacio Alzugaray and Elias Mueggler who shared their code for event-corner detection:

About

Implentation of the FA-Harris algorithm, described in Ruoxiang Li et al. "FA-Harris: A Fast and Asynchronous Corner Detector for Event Cameras", IROS 2019.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published