hector_imu_attitude_to_tf vs hector_localization

edit

Hi, I'm using an IMU and a laser scanner and would like to create a planar map, but estimate the pose with 6DoF, for example for handheld mapping.

As far as I can see, it is not possible to estimate the height (z-axis value) with the available sensors, hence 5DoF remain.

My question now is: According to the hector_mapping tutorial (here), it is possible to calculate the roll and pitch angles and broadcast the resulting transform between base_stabilized and base_link with hector_imu_attitude_to_tf. On the other hand, the package hector_pose_estimation within hector_localization apparently takes the pose estimate from hector_mapping on /poseupdate, fuses it with the IMU data and publishes the 6DoF pose.

So, are there any advantages/differences when taking hector_pose_estimation instead of hector_imu_attitude_to_tf for example for performing 2D handheld SLAM, or is it just the same? Is hector_pose_estimation maybe capable of doing something else that I miss (looking at the source code it seems to be pretty complex)?

Thanks, Philipp