Abstract
To compute reliable dense depth maps, a stereo algorithm must preserve depth discontinuities and avoid gross errors. In this paper, we show how simple and parallel techniques can be combined to achieve this goal and deal with complex real world scenes. Our algorithm relies on correlation followed by interpolation. During the correlation phase the two images play a symmetric role and we use a validity criterion for the matches that eliminate gross errors: at places where the images cannot be correlated reliably, due to lack of texture of occlusions for example, the algorithm does not produce wrong matches but a very sparse disparity map as opposed to a dense one when the correlation is successful. To generate a dense depth map, the information is then propagated across the featureless areas, but not across discontinuities, by an interpolation scheme that takes image grey levels into account to preserve image features. We show that our algorithm performs very well on difficult images such as faces and cluttered ground level scenes. Because all the algorithms described here are parallel and very regular they could be implemented in hardware and lead to extremely fast stereo systems.
Similar content being viewed by others
References
Anandan P (1989) A computational framework and an algorithm for the measurement of motion. Int J Computer Vision 2(3):283–310
Ayache N, Hansen C (1988) Rectification of images for binocular and trinocular Stereovision. Ninth International Conference on Pattern Recognition. Rome, Italy, November 1988, pp 11–16
Ayache N, Lustman F (1987) Fast and reliable passive trinocular Stereovision. First International Conference on Computer Vision
Barnard ST, Fischler MA (1982) Computational stereo. Comput Surv 14(4):553–572
Blake A, Zisserman A (1987) Visual Reconstruction. MIT Press, Cambridge, MA
Brown CM, Ballard DH (1982) Computer Vision. Prentice-Hall, Englewood Cliffs
Burt PJ, Yen C, Xu X (1982) Local correlation measures for motion analysis. In IEEE PRIP Conference, pp 269–274
Cailler C, Fornarix F-X, Heng P, Holtzer T (1990) Cocosun. Rapport de stage, Cerics
Faugeras OD (1988) A few steps toward artifical 3D vision. (Rapport de Recherche 790) INRIA, Sophia-Antipolis
Güelch E (1988) Results of test on image matching of isprs wg iii. Fourth Int Arch Photogrammetry Remote Sensing 27(III):254–271, and accompanying poster presentation
Hannah MJ (1988) Digital stereo image matching techniques. Int Arch Photogramrnetry Remote Sensing 27(III): 280–293
Hotz B (1991) Etude de techniques de stéréovision par corrélation. (Rapport des stage de dea) CNES, Toulouse, France
Kanade T, Okutomi M (1990) A stereo matching algorithm with an adaptative window: theory and experiment. Image Understanding Workshop
Mead C (1988) Analog vlsi for auditory and vision signal processing. INSPEC Conference, San Francisco, Calif
Meygret A, Thonnat M, Berthod M (1990) A pyramidal stereovision algorithm based on contour chain points. ECCV90 Conference, Antibes
Moravec H (1981) Robot visual navigation. UMI Research Press, Ann Arbor, MI
Mumford J, Shah D (1985) Boundary detection by minimizing functionals. CVPR85 Conference, San Francisco, Calif, June 1985, pp 22–28
Nishihara HK (1984) Practical real-time imaging stereo matcher. Optical Eng 23(5)
Nishihara HK, Poggio T (1983) Stereo vision for robotics. ISRR83 Conference, Bretton Woods, NH
Perona P, Malik J (1987) Scale space and edge detection using anisotropic diffusion. IEEE Computer Society Workshop on Computer Vision, Miami, Fl, pp 16–22
Poggio T, Torre V, Koch C (1985) Computational vision and regularization theory. Nature 317:314–319
Szeliski R (1989) Bayesian modeling of uncertainty in low-level vision. Kluwer Academic Press, Norwell MA
Szeliski R (1990) Fast surface interpolation using hierarchical basis functions. IEEE Trans Pattern Analysis Machine Intelligence 12(6):513–528
Terzopoulos D (1986) Image analysis multigrid relaxation methods. IEEE Trans Pattern Analysis Machine Intelligence 8(2): 129–139
Toscani G, Vaillant R, Deriche R, Faugeras OD (1989) Stereo camera calibration using the environment. 6th Scandinavian Conference on Image Analysis, pp 953–960
Author information
Authors and Affiliations
Additional information
This research was supported in part under the Centre National d'Etudes Spatiales VAP contract and in part under a Defence Advanced Research Projects Agency contract at SRI
Rights and permissions
About this article
Cite this article
Fua, P. A parallel stereo algorithm that produces dense depth maps and preserves image features. Machine Vis. Apps. 6, 35–49 (1993). https://doi.org/10.1007/BF01212430
Issue Date:
DOI: https://doi.org/10.1007/BF01212430