Video-based 3 D Reconstruction of Moving Scenes Using Multiple Stationary Cameras

In this paper, we describe a system for video-based 3D reconstruction of dynamic scenes using stereo techniques, with an eye to potential applications in human motion capture. We incorporate into our approach recent research results on stereo matching which make the system efficient and produce good-quality results. The implementation is built on top of Intel’s Open Source Computer Vision Library (OpenCV). Examples of 3D reconstruction results obtained from three synchronized video cameras are shown and discussed.

[1]  Milan Sonka,et al.  Image Processing, Analysis and Machine Vision , 1993, Springer US.

[2]  D. Scharstein,et al.  A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms , 2001, Proceedings IEEE Workshop on Stereo and Multi-Baseline Vision (SMBV 2001).

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

[4]  Pascal Fua,et al.  Articulated Soft Objects for Multiview Shape and Motion Capture , 2003, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  Carlo Tomasi,et al.  Depth Discontinuities by Pixel-to-Pixel Stereo , 1999, International Journal of Computer Vision.

[6]  S. P. Mudur,et al.  Three-dimensional computer vision: a geometric viewpoint , 1993 .