Generation of 3D Urban Model Using Cooperative Hybrid Stereo Matching

Abstract

We present a new hybrid stereo matching technique in terms of the co-operation of area-based stereo and feature-based stereo to build 3D site model from urban images. The core of our technique is that feature matching is carried out by the reference of the disparity evaluated by area-based stereo. Since the reference of the disparity can significantly reduce the number of feature matching combinations, feature matching error can be drastically minimized. One requirement of the disparity to be referenced is that it should be reliable to be used in feature matching. To measure the reliability of the disparity, in this paper, we employ the self-consistency of the disparity. Our suggested technique is applied to the detection of 3D line segments by 2D line matching using our hybrid stereo matching, which can be efficiently utilized in the generation of the rooftop model from urban images. Since occlusions are occurred around the outlines of buildings, we use multi-image stereo scheme by fusing 3D line segments extracted from several pairs of stereo images. The suggested method is evaluated on Avenches data set of Ascona aerial images. Experimental results indicate that the extracted 3D line segments have an average error of 0.5m and can be efficiently used to the construction of 3D site models using a simple 3D line grouping.