Title:Terrain reconstruction from oblique views

Author: Howard Schultz

Affiliation: Computer Vision Research Laboratory
             Department of Computer Science,
             University of Massachusetts, Box 34610 
             Amherst, MA 0l003-4610
Abstract:
  When a disparity map is computed from widely separated images 
  the perspective distortion may result in a large number of false 
  matches and poor reconstruction accuracy. This paper describes
  three image matching algorithms designed specifically to process 
  images taken from widely varying viewpoints. They include a new 
  match score, and modifications to standard subpixel and hierarchical 
  matching techniques.  The algorithms are incorporated into a 
  stereo analysis package and the system is tested by processing 
  a sequence of simulated images with base-to-height ratios that 
  varied between 0.25 and 2.25, and a single pair of high altitude 
  images with a base-to-height ratio of 0.63.  Analysis of the 
  simulated data showed that when these algorithms are implemented 
  the reconstruction accuracy remains independent of the 
  base-to-height ratio.

Keywords: Stereo vision, 3-D reconstruction, Terrain modeling.
          Image matching, Subpixel registration.