Dynamic Programming Approach for Semi-automated Road Extraction from Medium- and High-resolution Images

This paper presents a dynamic programming approach for semi-automated road extraction from mediumand high-resolution images. This method is a modified version of a pre-existent dynamic programming method for road extraction from low-resolution images. The basic assumption of this pre-existent method is that roads manifest as lines in low-resolution images and as such can be modelled and extracted as linear features. Contrary to this, roads manifest as ribbon features in mediumand high-resolution images and the goal of road extraction methods becomes the road centrelines. As a result, the original method can not accurately extract road centrelines from mediumand highresolution images. In view of this, we propose a modification of merit function of the original approach, which is carried out by a constraint function embedding road edge properties. Preliminary results demonstrate the modified algorithm’s potential in extracting road centrelines from mediumand high-resolution images. * Corresponding author.