Online Motion Estimation from Image and Inertial Measurements
Any type of content formally published in an academic journal, usually following a peer-review process.
We present two algorithms for estimating sensor motion from image and inertial measurements, which are suitable for use with inexpensive inertial sensors and in environments without known fiducials. The first algorithm is a batch method, which produces optimal estimates of the sensor motion, scene structure, and other parameters using measurements from the entire observation sequence simultaneously. The second algorithm recovers sensor motion, scene structure, and other parameters in an online manner, is suitable for use with long or “infinite” sequences, and handles sequences in which no feature is always visible.
We also describe initial results from running each algorithm on a sequence for which ground truth is available. We show that while image measurements alone are not sufficient for accurate motion estimation from this sequence, both batch and online estimation from image and inertial measurements produce accurate estimates of the sensors’ motion.