SfmLearner-Pytorch
SfmLearner-Pytorch copied to clipboard
Question about using oxts data
Thanks for great work! ` if scale is None: scale = np.cos(lat * np.pi / 180.)
pose_matrix = pose_from_oxts_packet(metadata[:6], scale) if origin is None: origin = pose_matrix odo_pose = imu2cam @ np.linalg.inv(origin) @ pose_matrix @ np.linalg.inv(imu2cam) ` I am curious about this part of the code which is not shown in the original sfmlearner. Could you tell me how this works(about math detail),or what book or reference should i read. Concretely
- what is the effect of the scale
- why ty = lat * np.pi * er / 180. different from the original implementation in https://github.com/utiasSTARS/pykitti
- the math detail about the transformation from (lat lon alt) to translation
- pose_matrix convert ? coordinate system to ? coordinate system,the effect of np.linalg.inv(origin) @ pose_matrix