SLAM_Dataset_ROSPKG

ROS package that collects and synchronizes data from following sensors: RGB-D Realsense camera, IMU and VICON motion camera system in order to create SLAM datasets. The package stores the data in a rosbag and it also can convert the rosbag in a SLAM dataset format, such as TUM dataset.
The synchorinization node works only if each topic have content.

Dependencies

This package was tested using:

1. ROS-melodic and Ubuntu18.04.
   
2. ROS-noetic and Ubuntu20.04.
   

Install Realsense camera package

install the realsense-viewer at : https://github.com/IntelRealSense/librealsense/blob/master/doc/distribution_linux.md

Run the following commands to install the dependencies:

sudo apt-get install ros-$ROS_DISTRO-realsense2-camera
sudo apt-get install ros-$ROS_DISTRO-openni-launch
mkdir rosbag image_dataset

After installed the dependencies and finish the compilation, make the following edits to the file:/opt/ros/melodic/share/realsense2_camera/launch/rs_rgbd.launch

1. edit the "depth_width", "depth_height" of the depth image for 640,480 respectively.
2. edit the "color_width", "color_height" of the rgb image for 640,480 respectively.
3. edit the "color_fps" of the image for 30
4. edit the "depth_fps" of the depth for 30

Run sensor synching executable

The synching executable can be run with the following command
On a terminal run:

roslaunch realsense2_camera rs_rgbd.launch

In a another terminal:

roslaunch slam_dataset_rospkg slam_rosbag_sync.launch

Create the rosbag

Once the synching executable is running, the following topics will be created

/camera/color/image_raw_sync
/camera/depth/image_rect_raw_sync
/imu/imu_sync
/vicon/pose_sync

Run the following command in another terminal to start recording the rosbag

rosbag record -O slam_dataset.bag /camera/color/image_raw_sync /camera/depth/image_rect_raw_sync /imu/imu_sync /vicon/pose_sync

Run SLAM image dataset executable

This executable can be used standalone to collect synched data from realsense RGB-D camera or can convert the image data from the rosbag. You might need to have the topic names at the src/collect_sync.cpp depending on your system. To run it standalone, run in a terminal the realsense camera launch:

roslaunch realsense2_camera rs_rgbd.launch

And on another terminal run the data collection package, if you don't provide the dataset path the system will store the data on the current folder:

rosrun slam_dataset_rospkg collect_sync _dataset_path:=[DST_DATASET_PATH]

To convert a rosbag, run the rosbag play command:

rosbag play <PATH TO ROSBAG>

And run the same command in another terminal:

rosrun slam_dataset_rospkg collect_sync _param:=[DST_DATASET_PATH]

Run image dataset executable with roslaunch

To run the package with camera visualization in parallel to the data acquisition, execute the following command:

roslaunch slam_dataset_rospkg slam_image_dataset.launch

Ground Truth extraction

To create a text file with the robots trajectory from a motion camera system, run the following command. You will need a terminal with roscore running and another terminal with your rosbag running as well. This will create a txt file with name ground_truth.txt

rosrun slam_dataset_rospkg gt_extractor _topic:=<topic Name>

Extract pose from T265

To create a text file with the robots trajectory from a RealSense T265, run the following command. You will need a terminal with roscore running and another terminal with your rosbag running as well. This will create a txt file with name t265_odom.txt

rosrun slam_dataset_rospkg t265_odom _topic:=<topic Name>