Transform functions status

[elemoine]

The goal of this issue is to provide some status on the transform functions. With the recent work on the integration of the Stereopolis La Villette data we've introduced a few things that may be Stereopolis-specific. This issue provides a list of these things.

1. In pointcloud the li_patch_spherical_to_cartesian function adds π/2 to Φ. This is related to the Φ angle in Stereopolis EPT files.
2. In cli-li3ds the read_lidar_rigid_transform_params function of import-platform command uses -HeadingShift instead of HeadingShift. This was related to some email from Julien R Haridon, with information on the way he uses the values in the TMS calibration file (TerraMatch).
3. In cli-li3ds the read_lidar_rigid_transform_params function of import-platform command applies rotations around y and z to take into account that the axis of the INS and LiDAR referentials are not aligned. The images of this IGN paper were used to determine the positions/orientations of the LiDAR relative to the INS.
4. In cli-li3ds the import-platform applies a rotation by π around Z between the INS and the camera group. This is to take into account the orientation/position of the camera group relative to the INS.
5. In api-li3ds the foreignpc.ForeignViews class makes assumptions on the heading, roll, pitch angles found in Sbet files. I am not sure whether all Sbet files use the same angle conventions, or if the convention used here is Stereoplis-specific.

@mbredif please tell us if you want to discuss all this further.