Hello,
I tried to run FAST-LIO in ROs 2 humble in Ubuntu 22.04 following the instructions in the readme.
ros2 launch fast_lio mapping.launch.py config_file:=avia.yaml
ros2 bag play outdoor_Mainbuilding_100Hz_2020-12-24-16-46-29
I used rosbags to convert the ROS1 bag file indicated in the readme file to ROS2 bag file.
The message type of the converted livox/lidar topic was livox_ros_driver/msg/CustomMsg and I manually changed the message type to livox_ros_driver2/msg/CustomMsg both in the metadata.yaml and xyz.db3 database table using sqlite3 command.
However, the generated map by FAST-LIO with this converted bag file is not as accurate as the one generated with ROS1, and also the points in the map keep being deleted after some time ultimately leaving nothing(until there's no map) as shown in the GIF below.
For more detailed visualization, I put video for both ros1 and ros2 here.
In section V of the paper it states that all points in the subtraction area between the new map region and the old one will be deleted from the ikd-Tree by a box-wise delete operation. I am not sure if it's this deletion operation that causes an empty map ultimately. If it is, how can I persist the map without any points being deleted from it? If not, why are the points continuously being deleted until the map is eventually empty?
Hello,
I tried to run FAST-LIO in ROs 2 humble in Ubuntu 22.04 following the instructions in the readme.
ros2 launch fast_lio mapping.launch.py config_file:=avia.yamlros2 bag play outdoor_Mainbuilding_100Hz_2020-12-24-16-46-29I used rosbags to convert the ROS1 bag file indicated in the readme file to ROS2 bag file.
The message type of the converted
livox/lidartopic waslivox_ros_driver/msg/CustomMsgand I manually changed the message type tolivox_ros_driver2/msg/CustomMsgboth in themetadata.yamlandxyz.db3database table using sqlite3 command.However, the generated map by FAST-LIO with this converted bag file is not as accurate as the one generated with ROS1, and also the points in the map keep being deleted after some time ultimately leaving nothing(until there's no map) as shown in the GIF below.
For more detailed visualization, I put video for both ros1 and ros2 here.
In section V of the paper it states that all points in the subtraction area between the new map region and the old one will be deleted from the ikd-Tree by a box-wise delete operation. I am not sure if it's this deletion operation that causes an empty map ultimately. If it is, how can I persist the map without any points being deleted from it? If not, why are the points continuously being deleted until the map is eventually empty?