The sequence is first traversed and separated in subsequences.
For subsequences with sufficient 3D structure (case A) the approach described in Section 5 is followed so that the projective structure and motion is recovered. When a triplet corresponds to case B, only planar features are tracked and reconstructed (in 2D). A possible partitioning of an image sequence is given in Table 5.2. Note that the triplet 3-4-5 would cause an approach based on [155] to fail.
Suppose the plane is labeled as a dominant plane from view based on features tracked in views . In general, some feature points
located on will have been reconstructed in 3D from previous views (e.g. and ). Therefore, the coefficients of can be computed from
. Define
as the right null space of ( matrix).
represents 3 supporting points for the plane and let
be the corresponding image projections. Define the homography
, then the 3D reconstruction of image points located in the plane are obtained as follows:
(E11) |
(E12) |