Camera Log HDF5 Format

The executable tricamera_log_to_hdf5 can be used to convert existing logs of TriCamera data with object pose information (as produced by this package) to the more portable HDF5 format.

The data format is the same as described in TriCameraLogger HDF5 Format but extended by the following datasets:

• object_pose/has_pose: Boolean indicating whether the object was found at all. If false, confidence is expected to be zero and values of position and orientation are meaningless.

• object_pose/position: The estimated x, y, z-position of the object w.r.t. the world frame. Dimensions: (n_observations, 3).

• object_pose/orientation: The estimated orientation of the object w.r.t. the world frame. Given as quaternions (x, y, z, w). Dimensions: (n_observations, 4).

• object_pose/confidence: The confidence measure of the pose estimation. Dimensions: (n_observations, 1).