I’m not expert in issues related with positional tracking. But, Can i get pose information of positional tracking from a recorded svo2 file?
There is a method on python, get_pose(), that gets a Transform class. From this structure I get an orientation data (quaternion), and translation, but this last data doesn’t change (always is 0,0,0).
Any idea what could be going on? Thanks
Hi @benferpo
you can use an SVO2 as input for an application ZED SDK-based.
The information stored in the SVO2 (raw images and raw sensors data) will be used as if the camera used to record it were connected to the PC running the application.
I was finally able to get “two” pose information from the svo2 file:
I get a first position, let’s say 1, which is calculated only from the IMU data zed_imu.get_pose()
Another position 2 that also uses visual corrections and SLAM technology.
camera_pose = sl.Pose()
tracking_state = zed.get_position(camera_pose,sl.REFERENCE_FRAME.CAMERA)
But I still don’t know why the first pose (which only uses IMU data) the translation values are always 0. I took the capture in a vehicle where the translation is expected to change (and so it does for pose 2)