A Hybrid 3DoF Pose Estimation Method based on Camera and Lidar Data

This repository implements a hybrid 3DoF pose estimation method combining RGB camera and single-channel Lidar data. The detailed introduction can be found in ＂A Hybrid 3DoF Pose Estimation Method based on Camera and Lidar Data＂, ROBIO 2017.

Overview

This repository is based on ROS and includes three packages, where library contains driver and basic process for Hokuyo urg-04lx-ug01 Lidar, lidar_based_method and vision_based_method estimate targets' pose using Lidar and camera, respectively.

Prerequisites

Since ROS is utilized to subscribe data, please make sure you have installed ROS indigo/kinetic.

Installing

After you have download this repository, make the workspace at the root directory using

catkin_make

And source

source ./devel/setup.bash

Running

After you successfully make all codes, you can run the code using (make sure you have pluged in the camera and Lidar)

roslaunch lidar_based_method lidar.launch

roslaunch vision_based_method vision.launch

Now you can check the rostopic list

rostopic list

where "block_single_location" and "camera_pose" give the estimation from lidar and camera, respectively. And now you can further fuse these two estimation to get a more robust result.

Noted

If you use this code for academic purposes, please cite it as:

@inproceedings{Tingguang:2017,
  author          = {Tingguang Li, Delong Zhu, and Max Q.-H. Meng},
  title           = {A Hybrid 3DoF Pose Estimation Method based Camera and Lidar Data},
  booktitle       = {Robotics and Biomimetics (ROBIO), 2017 IEEE International Conference on},
  year            = {2017},
  organization      ={IEEE}
}