HD-CCSOM: Hierarchical and Dense Collaborative Continuous Semantic Occupancy Mapping through Label Diffusion
This is a novel collaborative continuous semantic occupancy Mapping algorithm, opening up a new field of mapping.
catkin_ws/src$ git clone https://github.com/BIT-DYN/HD-CCSOM
catkin_ws/src$ cd ..
catkin_ws$ catkin_make
catkin_ws$ source ~/catkin_ws/devel/setup.bash
catkin_ws$ source /opt/intel/compilers_and_libraries/linux/bin/compilervars.sh intel64
catkin_ws$ catkin_make -DCMAKE_C_COMPILER=icc -DCMAKE_CXX_COMPILER=icpc
catkin_ws$ source ~/catkin_ws/devel/setup.bash
Download the demo dataset and place them in directory "catkin_ws/src/HD-CCSOM/data/*.bag".
First, each robot builds a local semantic map.
catkin_ws$ roslaunch hd_ccsom multi_gazebo.launch
Then, you can run the command through the new terminal to complete the map sharing and fusion.
rosservice call /robot2/manual_triger
Each robot stores a local map and a global map.
@inproceedings{deng2022hd,
title={Hd-ccsom: Hierarchical and dense collaborative continuous semantic occupancy mapping through label diffusion},
author={Deng, Yinan and Wang, Meiling and Yang, Yi and Yue, Yufeng},
booktitle={2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
pages={2417--2422},
year={2022},
organization={IEEE}
}