Related Publications:
@inproceedings{Yang2017Real,
title={Real-time Globally Consistent 3D Grid Mapping},
author={Yang, Sining and Yang, Shaowu and Yi, Xiaodong},
booktitle={IEEE International Conference on Robotics and Biomimetics},
year={2017},
}
Dependencies
1、OpenCV(http://docs.opencv.org/master/d9/df8/tutorial_root.html)
2、The modified octomap(https://github.com/daysun/octomap/) && octovis(http://wiki.ros.org/octovis)
3、PCL(http://pointclouds.org/downloads/)
4、genmsg(https://github.com/ros/genmsg)
5、octomap_msgs(https://github.com/OctoMap/octomap_msgs/tree/indigo-devel)
6、the modified orb-slam2(https://github.com/daysun/ORB-SLAM2)
Compile
You should put this project under your ROS work space, such as your_ros_path/src/
cd your_ros_path
catkin_make
Usage
1、 rosrun octomap_ros rcv_test
Before executing this, make sure ORB-SLAM2(https://github.com/daysun/ORB-SLAM2) are running.
The experimental rosbag data can be attained in https://pan.baidu.com/s/1boC6X7h, and the corresponding parameters can be attained in /data/TUM1.yaml.
This procedure receives msg(Type: Id_PointCloud2) from ROS, which named “/ORB_SLAM/pointcloud2” . And it will transform the msg into the octomap format.
Try "Ctrl+C" to stop it any time you want, which would generate origin.ot. You can view it in octovis.
octovis origin.ot
2、 rosrun octomap_ros runtime_test
For experiments.
Show that our system can perform on-line. The usage is the same with the former one.
And the results built by original grid mapping, original grid mapping and local grid mapping, and complete update strategy using our experimental data can be attained in /data/*.ot.
The experiments are also shown in a vedio online(http://v.youku.com/v_show/id_XMjg5MTA1OTg4OA).