Using disparity functionals for stereo correspondence and reconstruction
Jason D. Eastman · 1990
In this dissertation, we investigate techniques in computer vision for stereo matching and calibration. We investigate constraints for stereo matching that derive from an analytic model of surface depth. We formulate stereo as a single stage process in which potential feature point or contour matches interact to provide support for local estimates of a polynomial model of disparity, the disparity functional. We present an algorithm that integrates the disparity functional with multiresolution matching of zero-crossings to derive depth to surface patches. The analyticity of the disparity field is thereby exploited early in the matching process, and yields surface reconstruction as a direct byproduct of correspondence. In addition, we investigate practical difficulties in the calibration of a two camera stereo system in an uncontrolled environment for the case where the relative orientation angles are small and the distance between the two cameras is known. This is done by deriving explicit analytical solutions for the relative pan, tilt and roll angles in terms of the world pan angle (often referred to as gaze angle) and the coordinates of the image points used in their computation. We show that the sensitivity of the computation of the relative pan and roll angles to numeric error greatly depends on the choice of image points used for the computation of these angles, whereas the sensitivity of the computation of the relative tilt angle to the error due to image center position is only marginally affected by our choice of the image points.