In this paper, we propose a new 3D reconstruction method for general 3D planar curves based on curve correspondences on two views. By fitting the measured and transferred points using spline curves and minimizing the 2D Euclidean distance from measured and transferred points to fitted curves, we obtained an optimum homography which relates the curves across two views. Once two or more homographies are computed, 3D projective reconstruction of those curves can be readily performed. The method offers the flexibility to reconstruct 3D planar curves without the need of point-to-point correspondences, and deals with curve occlusions automatically. © 2010 Springer-Verlag.
CITATION STYLE
Zhang, X. B., Tang, A. W. K., & Hung, Y. S. (2010). Projective reconstruction of general 3D planar curves from uncalibrated cameras. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 6454 LNCS, pp. 21–30). https://doi.org/10.1007/978-3-642-17274-8_3
Mendeley helps you to discover research relevant for your work.