DIRECT RECONSTRUCTION OF PLANAR SURFACES BY STEREO VISION

This paper studies the problem of reconstructing a planar surface from stereo images of multiple feature points that are known to be coplana r in the scene. We present a direct method by applying maximum likelih ood estimation based on a statistical model of image noise. The signif icant fact about our method is that not only the 3-D position of the s urface is reconstructed accurately but its reliability is also compute d quantitatively. The effectiveness of our method is demonstrated by d oing numerical simulation.