论文信息 - VRML model generation from disparity field

VRML model generation from disparity field

We present an approach to obtain a VRML tex-tured model of a static scene from 2 views extracted from an uncalibrated sequence. This approach is based on a 2d triangulation of an image of the sequence and its corresponding disparity eld. A 3d triangulation of the scene is then obtained from this 2d triangulation and a weak calibration. A geometric correction of the extracted image allows textures extraction which is necessary for a realistic model. The proposed technique has been tested with real and synthetic images.

Lionel Oisel | Luce Morin | Franck Galpin

[1] Richard I. Hartley,et al. Estimation of Relative Camera Positions for Uncalibrated Cameras , 1992, ECCV.

[2] Stéphane Pateux. Segmentation spatio-temporelle et codage oriente-regions de sequences video bases sur le formalisme mdl , 1998 .

[3] Lionel Oisel. Reconstruction 3d de scenes complexes a partir de sequences video non calibrees : estimation et maillage d'un champ de disparite , 1998 .

[4] Thomas S. Huang,et al. Uniqueness and Estimation of Three-Dimensional Motion Parameters of Rigid Objects with Curved Surfaces , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5] Olivier D. Faugeras,et al. Relative 3D positioning and 3D convex hull computation from a weakly calibrated stereo pair , 1993, 1993 (4th) International Conference on Computer Vision.