GENERATING SEAMLESS STEREO MOSAICS FROM AERIAL VIDEO

This paper is directed at the development of tools for environmental science and other aerial video ap plications requiring the mapping of video into absolute world coordinates. As part of this effort, we propose a new method that automatically and efficiently generates seamless st ereo mosaics from an aerial video camera moving ove r a large distance. An important part of our approach is a pa rallel-perspective mosaic representation that can w ell support large seamless mosaics and also can capture the inh erent 3D information. We will show that the conditi on under which such stereo mosaics can be created is a rathe r general motion (3D rotation and translation) with a dominant direction, and the parallel-perspective stereo mosa ics can be converted into a digital elevation map (DEM) after 3D recovery. A novel method PRISM (parallel-perspectiv e ray interpolation for stereo mosaicing) is propos ed to generate dense and seamless parallel-perspective st ereomosaics from sparse image sequences with motion par allax. The advantages of stereo mosaics are shown in repre sentation, computation, 3D recovery and 3D visualiz ation. Issues on the implementations of geo-references mos aicing as well as free mosaicing are discussed. Exp erimental results on several long video sequences are given.

[1]  A.H. Haddad,et al.  Applied optimal estimation , 1976, Proceedings of the IEEE.

[2]  Edward H. Adelson,et al.  A multiresolution spline with application to image mosaics , 1983, TOGS.

[3]  鄭絳宇,et al.  Panoramic Representation for Route Recognition by a Mobile Robot , 1990 .

[4]  Hiroshi Ishiguro,et al.  Omni-directional stereo for making global map , 1990, [1990] Proceedings Third International Conference on Computer Vision.

[5]  Rama Chellappa,et al.  Estimating the Kinematics and Structure of a Rigid Object from a Sequence of Monocular Images , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  Harpreet S. Sawhney.  Simplifying motion and structure analysis using planar parallax and image warping , 1994, Proceedings of 12th International Conference on Pattern Recognition.

[7]  Steven M. Seitz,et al.  Physically-valid view synthesis by image interpolation , 1995, Proceedings IEEE Workshop on Representation of Visual Scenes (In Conjunction with ICCV'95).

[8]  Richard Szeliski,et al.  Direct methods for visual scene reconstruction , 1995, Proceedings IEEE Workshop on Representation of Visual Scenes (In Conjunction with ICCV'95).

[9]  Shenchang Eric Chen,et al.  QuickTime VR: an image-based approach to virtual environment navigation , 1995, SIGGRAPH.

[10]  Harpreet S. Sawhney,et al.  Registration of video to geo-referenced imagery , 1998, Proceedings. Fourteenth International Conference on Pattern Recognition (Cat. No.98EX170).

[11]  Heung-Yeung Shum,et al.  Panoramic Image Mosaics , 1998 .

[12]  Harry Shum,et al.  Construction and refinement of panoramic mosaics with global and local alignment , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[13]  Andrew W. Fitzgibbon,et al.  Bundle Adjustment - A Modern Synthesis , 1999, Workshop on Vision Algorithms.

[14]  Shmuel Peleg,et al.  Stereo panorama with a single camera , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[15]  Allen R. Hanson,et al.  STEREO MOSAICS FROM A MOVING VIDEO CAMERA FOR ENVIRONMENTAL MONITORING , 1999 .