Stereo map surface calculus optimization using radial basis functions neural network interpolation

Abstract

Matching two points in distinguished images is one of the actual challenges in digital image processing. The currents techniques of stereo imaging are not ideal and slow, not offering a valuable solution for practical scenarios like real time robotics vision or vehicles navigation. This paper will presents a approach for optimization of disparity calculus, introducing a sparse matching of stereo disparity maps and surface reconstruction by RBFs interpolation of empty spaces. © 2009 Springer-Verlag Berlin Heidelberg.