Adds two-view triangulation capability. Triangulation is based on the direct linear transform (an algebraic method). One can supply the fundamental matrix which will result in a projective reconstruction. Alternatively, If calibration matrices are known, then one can triangulate using the essential matrix. This will result in an Euclidean reconstruction up to an unknown scale factor. One can also triangulate by directly supplying a pair of projection matrices.
This commit also adds Kanatani's optimal correction so that a set of noisy corresponding points satisfy the epipolar constraint. However, the triangulation method currently does not use this correction.
Adds two-view triangulation capability. Triangulation is based on the direct linear transform (an algebraic method). One can supply the fundamental matrix which will result in a projective reconstruction. Alternatively, If calibration matrices are known, then one can triangulate using the essential matrix. This will result in an Euclidean reconstruction up to an unknown scale factor. One can also triangulate by directly supplying a pair of projection matrices.
This commit also adds Kanatani's optimal correction so that a set of noisy corresponding points satisfy the epipolar constraint. However, the triangulation method currently does not use this correction.