briannemsick / radmap_point_clouds

Preprocessing, coordinate frame calibration, configuration files, and launching procedure used to generate point clouds with Google Cartographer for the RadMAP acquisition system. The RadMAP acquisition system consists of two LIDARS, differential GPS, two Ladybug 360 cameras, and an IMU.

The radmap_point_clouds packages does the following:

1. converts velodyne_msgs/VelodyneScan --> sensor_msgs/PointCloud2

2. merges {sensor_msgs/PointCloud2,sensor_msgs/Imu,sensor_msgs/NavSatFix}

3. Aligns 2 velodynes in SE3 (rotation and translation)

4. Provides {.lua,.launch,.urdf} configuration files for cartographer

Notes:

1. See installation.txt for step by step guide to build this package

2. The bagfile names are assumed to be {port,starboard,imu}.bag with the appropriate topic names {/novatel_fix,/novatel_imu,/velodyne_packets_port, /velodyne_packets_starboard}

Requirements:

1. https://github.com/ros-drivers/velodyne

2. https://github.com/googlecartographer/cartographer_ros

3. http://www.cloudcompare.org/release/

Recommendations (add to .bashrc):

1. source /home/[user]/radmap/internal/devel/setup.bash

2. export RADMAP="/home/[user]/radmap/internal/src/radmap_point_clouds"

3. export RADBAGS="/home/[user]/radmap/internal/src/radmap_point_clouds/bagfiles"

Merging the bags

1. Place port.bag, starboard.bag, imu.bag into [folder]
2. cd [folder]
3. roslaunch radmap_point_clouds lasers.launch
4. rosrun radmap_point_clouds lasers
5. python $RADMAP/src/bag_merging/bag_merge_radmap.py

Run cartographer

1. copy params/radmap_config to cartographer/[dev branch]/share/cartographer_ros/radmap_config
2. roslaunch cartographer_ros demo_radmap.launch bag_filename:=${RADBAGS}/[bagfile]
3. rosservice call /finish_trajectory [name]
4. roslaunch cartographer_ros assets_writer_radmap.launch bag_filenames:=${RADBAGS}/[bagfile] trajectory_filename:=/home/[user]/.ros/[name].pb
5. cd /home/[user]/ --> CloudCompare --> visualize points.ply

(Optional) Vizualize aligned scans

1. roslaunch radmap_point_clouds calibrated_tf.launch
2. rosbag play [bagfile]
3. rviz -d $RADMAP/.rviz/aligned_scans.rviz

(Optional) Coordinate frame calibration

1. velodyne_calib.m
2. fill in calibrated_tf.launch
3. roslaunch radmap_point_clouds calibrated_tf.launch
4. rosrun tf tf_echo imu_link velodyne_port --> radmap.lau - velodyne_port_joint - origin
5. rosrun tf tf_echo imu_link velodyne_starboard --> radmap.lau - velodyne_starboard_joint - origin