Object-Space Road Extraction in Rural Areas Using Stereoscopic Aerial Images

This letter proposes a semiautomatic method for 3-D road extraction in rural areas using stereoscopic aerial images. A strategy based on the dynamic programming algorithm provides a solution to the road extraction problem in the object space. The direct tracing of road centerlines in the object space necessitates mathematical relationships connecting road points in the stereoscopic image spaces and in the object space, enabling integration of radiometric information from the stereoscopic images into the associated energy function. The extraction process begins by first measuring a few seed points in one image of the stereoscopic pair and then transforming these into the object-space reference system. Experimental results show that the proposed method was efficient and usually provided accurate road centerlines.