Fig 1.
Angular direction of pixel p.
Fig 2.
Flow chart of the 3D reconstruction method.
Fig 3.
Triangulation.
Fig 4.
Flow chart of the projection-based Delaunay triangulation.
(a) 3D point set (b) plane point set (c) plane triangulation (d) 3D triangulation (e) texture mapping.
Fig 5.
Flow chart of the Bowyer-Watson algorithm.
(a) inserting the point P (b) searching for the influence triangles (c) deleting common edge AB (d) forming triangles.
Fig 6.
Method of quickly locating insertion point based on direction search.
Fig 7.
3D reconstruction experiment of scene Cloth1.
(a) left image, (b) right image, (c) disparity image, (d) Delaunay triangulation, (e) view 1, (f) view 2, (g) view 3, (h) view 4. (a) and (b) are republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Fig 8.
3D reconstruction experiment of scene Wood1.
(a) left image, (b) right image, (c) disparity image, (d) Delaunay triangulation, (e) view 1, (f) view 2, (g) view 3, (h) view 4. (a) and (b) are republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Fig 9.
3D reconstruction experiment of scene Djembe.
(a) left image, (b) right image, (c) disparity image, (d) Delaunay triangulation, (e) view 1, (f) view 2, (g) view 3, (h) view 4. (a) and (b) are republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Fig 10.
3D reconstruction experiment of scene Piano.
(a) left image, (b) right image, (c) disparity image, (d) Delaunay triangulation, (e) view 1, (f) view 2, (g) view 3, (h) view 4. (a) and (b) are republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Fig 11.
3D reconstruction experiment of scene Shelves.
(a) left image, (b) right image, (c) disparity image, (d) Delaunay triangulation, (e) view 1, (f) view 2, (g) view 3, (h) view 4. (a) and (b) are republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Fig 12.
Wood1 scene in the Middlebury2 dataset.
The left image in the figure is republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Fig 13.
Piano scene and the Shelves scene in the Middlebury3 dataset.
The left image in the figure is republished from [vision.middlebury.edu/stereo/] under a CC BY license, with permission from [D. Scharstein], original copyright [2002].
Table 1.
Comparison of 3D reconstruction errors for six scenes in Middlebury2 and Middlebury3 (measurement unit: mm).
Table 2.
Depth information of 3D reconstruction in the KITTI 2012 benchmark.
Table 3.
Depth information of 3D reconstruction in the KITTI 2015 benchmark.
Fig 14.
3D reconstruction experiment of scenes in KITTI 2012.
It can be downloaded via https://osf.io/7srj4.
Fig 15.
3D reconstruction experiment of scenes in KITTI 2015.
It can be downloaded via https://osf.io/9kzma.
Table 4.
Results of Delaunay triangulation based on fast point positioning (measurement unit: ms).