Pose Estimation for Vehicle-mounted Cameras via Horizontal and Vertical Planes

We propose two novel solvers for estimating the egomotion of a calibrated\ncamera mounted to a moving vehicle from a single affine correspondence via\nrecovering special homographies. For the first class of solvers, the sought\nplane is expected to be perpendicular to one of the camera axes. For the second\nclass, the plane is orthogonal to the ground with unknown normal, e.g., it is a\nbuilding facade. Both methods are solved via a linear system with a small\ncoefficient matrix, thus, being extremely efficient. Both the minimal and\nover-determined cases can be solved by the proposed methods. They are tested on\nsynthetic data and on publicly available real-world datasets. The novel methods\nare more accurate or comparable to the traditional algorithms and are faster\nwhen included in state of the art robust estimators.\n