This paper has the purpose to show a stabilization video-streaming methodology feasible to low-power wearable devices. Thanks to an Inertial Measurement Unit (IMU) mounted together with the camera we are ready to stabilize directly on video-stream without the delay and the complexity due to image processing used by classic software stabilization techniques. The IMU gives information about the angle rotation respect to the three main orthogonal axes of the camera; the wearable device transmits the video along with the IMU data synchronized frame per frame then a base station receives and stabilizes while renders the video. The result is that the shaking and the unwanted motions of the human body wearing the system are compensated giving a clear and stable video. Numeric results prove that the video is more stable: we cut-off the half of the motion noise in the scene.
A real-time video stream stabilization system using inertial sensor
Longobardi A.;Tecchia F.;Carrozzino M.;Bergamasco M.
2019-01-01
Abstract
This paper has the purpose to show a stabilization video-streaming methodology feasible to low-power wearable devices. Thanks to an Inertial Measurement Unit (IMU) mounted together with the camera we are ready to stabilize directly on video-stream without the delay and the complexity due to image processing used by classic software stabilization techniques. The IMU gives information about the angle rotation respect to the three main orthogonal axes of the camera; the wearable device transmits the video along with the IMU data synchronized frame per frame then a base station receives and stabilizes while renders the video. The result is that the shaking and the unwanted motions of the human body wearing the system are compensated giving a clear and stable video. Numeric results prove that the video is more stable: we cut-off the half of the motion noise in the scene.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.