Multi-view Orientation Estimation using Bingham Mixture Models

Publication Type:

Conference Paper


IEEE International Conference on Automation, Quality and Testing, Robotics (2016)


This paper describes a multi-view pose estimation system, that is exploiting the mobility of a depth sensor through mounting it onto a robotic manipulator. Given a pose estimation algorithm that performs feature extraction and matching to a model database, we investigate the probabilistic modeling of the pose space as well as the measurement uncertainty, to be used in a sequential state estimation approach. Uncertainties in the position in Euclidean space can be modeled by 3d Gaussians, but the space of rotations in 3d, the special orthogonal group SO(3), requires approaches from directional statistics. A convenient representation is to represent orientations by unit quaternions, and then use the Bingham distribution on the 4d sphere they define. This has the advantage of modeling the symmetry in the quaternion representation correctly, and to leave degrees of freedom unconstrained (which is especially useful if an object is rotationally symmetric, with no unique quaternion describing its orientation). In our experiments we test different sequential fusion methods, optimize their parameters, and investigate how does the derived filter perform in a case with high uncertainties.

File Attach: