Determining 3D structure and motion of man-made objects from image corners

This paper deals with motion estimation from image corner correspondences for two cases: the orthogonal corner and the general corner with known space angles. The contribution of the paper is threefold. First, the three dimensional structure of a corner is recovered easily from its image by introducing a new coordinate system. Second, it is shown that the correspondence of one corner and two points over two views is sufficient to determine the motion uniquely, i.e., the rotation and translation. Third, experiments conducted using both simulated data and real images present good results.