Dramatic Improvements to Feature Based Stereo

The camera registration extracted from feature based stereo is usually considered sufficient to accurately localize the 3D points. However, for natural scenes the feature localization is not as precise as in man-made environments. This results in small camera registration errors. We show that even very small registration errors result in large errors in dense surface reconstruction.We describe a method for registering entire images to the inaccurate surface model. This gives small, but crucially important improvements to the camera parameters. The new registration gives dramatically better dense surface reconstruction.

[1]  Olivier Faugeras,et al.  Three-Dimensional Computer Vision , 1993 .

[2]  David A. Maluf,et al.  Bayesian super-resolved surface reconstruction from images , 2000, Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662).

[3]  Edwin E. Catmull,et al.  A hidden-surface algorithm with anti-aliasing , 1978, SIGGRAPH.

[4]  Christopher G. Harris,et al.  A Combined Corner and Edge Detector , 1988, Alvey Vision Conference.

[5]  Zhengyou Zhang,et al.  A Flexible New Technique for Camera Calibration , 2000, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  O. Faugeras Three-dimensional computer vision: a geometric viewpoint , 1993 .

[7]  Pascal Fua,et al.  Registration without correspondences , 1994, 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[8]  M. Carter Computer graphics: Principles and practice , 1997 .

[9]  Daniel Snow,et al.  Determining Generative Models of Objects Under Varying Illumination: Shape and Albedo from Multiple Images Using SVD and Integrability , 1999, International Journal of Computer Vision.

[10]  V. Leitáo,et al.  Computer Graphics: Principles and Practice , 1995 .

[11]  Peter Cheeseman,et al.  (Almost) Featureless Stereo: Calibration and Dense 3D Reconstruction Using Whole Image Operations , 2001 .

[12]  Peter R. Atherton,et al.  Hidden surface removal using polygon area sorting , 1977, SIGGRAPH.

[13]  Rachid Deriche,et al.  A Robust Technique for Matching two Uncalibrated Images Through the Recovery of the Unknown Epipolar Geometry , 1995, Artif. Intell..