3D solutions to the aperture problem
The motion of image edge detectors is used to infer 3D scene information by simultaneously solving the aperture and structure-from-motion problems. It is shown how additional constraints or information are required and two formal solutions are given. In one the motion of edge features at eight or mo...
| Main Authors: | , , , |
|---|---|
| Format: | Conference item |
| Published: |
European Association for Artificial Intelligence
1984
|
| _version_ | 1797079458351415296 |
|---|---|
| author | Buxton, B Buxton, H Murray, D Williams, N |
| author_facet | Buxton, B Buxton, H Murray, D Williams, N |
| author_sort | Buxton, B |
| collection | OXFORD |
| description | The motion of image edge detectors is used to infer 3D scene information by simultaneously solving the aperture and structure-from-motion problems. It is shown how additional constraints or information are required and two formal solutions are given. In one the motion of edge features at eight or more image points is used to solve for the motion and surface parameters (the 8-point algorithm). In the other it is assumed that the motion is known as a priori (the 3-point algorithm). This algorithm is used to illustrate the numerical stability of these solutions. |
| first_indexed | 2024-03-07T00:46:09Z |
| format | Conference item |
| id | oxford-uuid:84b9cdb2-332d-4d79-92ed-2869ed274e96 |
| institution | University of Oxford |
| last_indexed | 2024-03-07T00:46:09Z |
| publishDate | 1984 |
| publisher | European Association for Artificial Intelligence |
| record_format | dspace |
| spelling | oxford-uuid:84b9cdb2-332d-4d79-92ed-2869ed274e962022-03-26T21:52:58Z3D solutions to the aperture problemConference itemhttp://purl.org/coar/resource_type/c_5794uuid:84b9cdb2-332d-4d79-92ed-2869ed274e96Symplectic Elements at OxfordEuropean Association for Artificial Intelligence1984Buxton, BBuxton, HMurray, DWilliams, NThe motion of image edge detectors is used to infer 3D scene information by simultaneously solving the aperture and structure-from-motion problems. It is shown how additional constraints or information are required and two formal solutions are given. In one the motion of edge features at eight or more image points is used to solve for the motion and surface parameters (the 8-point algorithm). In the other it is assumed that the motion is known as a priori (the 3-point algorithm). This algorithm is used to illustrate the numerical stability of these solutions. |
| spellingShingle | Buxton, B Buxton, H Murray, D Williams, N 3D solutions to the aperture problem |
| title | 3D solutions to the aperture problem |
| title_full | 3D solutions to the aperture problem |
| title_fullStr | 3D solutions to the aperture problem |
| title_full_unstemmed | 3D solutions to the aperture problem |
| title_short | 3D solutions to the aperture problem |
| title_sort | 3d solutions to the aperture problem |
| work_keys_str_mv | AT buxtonb 3dsolutionstotheapertureproblem AT buxtonh 3dsolutionstotheapertureproblem AT murrayd 3dsolutionstotheapertureproblem AT williamsn 3dsolutionstotheapertureproblem |