Introduction

The omt_coscan package is designed for the collaborative scanning of unknown environments driven by Optimal Mass Transport(OMT) using multiple RGBD cameras. The key idea is to formulate the collaborative scanning task as a resource distribution problem. We solve the problem by a tailored OMT algorithm to achieve high scanning coverage while keeping low scanning effort.

Overview

The package consists of three sub-packages:

1) co_scan for the core algorithm

2) srbot_description for robot(RGBD sensor) params

3) virtual_scan for the simulator of virtual scanning using Gazebo.

Contributors and Publication

The contributors to this package are Siyan Dong(siyandong.3@gmail.com), Kai Xu(kevin.kai.xu@gmail.com) and Qiang Zhou(bamboosdu@gmail.com). This method has been published at SIGGRAPH 2019. Please refer to https://kevinkaixu.net/projects/multirobot.html. If you find the package helpful for your research, please consider citing the following paper:

@article{dong2019multi,
  title={Multi-robot collaborative dense scene reconstruction},
  author={Dong, Siyan and Xu, Kai and Zhou, Qiang and Tagliasacchi, Andrea and Xin, Shiqing and Nie{\ss}ner, Matthias and Chen, Baoquan},
  journal={ACM Transactions on Graphics (TOG)},
  volume={38},
  number={4},
  pages={84},
  year={2019},
  publisher={ACM}
}

Download and Install

Dependencies

This package depends on OpenCV, CGAL and OctoMap. Please install these libraries first.

From Source

Please download the source code:

catkin_ws/src$ git clone https://github.com/siyandong/CoScan.git

Modify the direcrotries of OpenCV, CGAL and OctoMap in CMakeLists.txt. Then

catkin_ws$ catkin_make_isolated

Run

catkin_ws$ roslaunch virtual_scan simulator.launch
catkin_ws$ roslaunch virtual_scan data_sever.launch
catkin_ws$ rosrun co_scan co_scan