Semi-automatic road extraction by dynamic programming optimization in the object space: single image case

• Published in: Revista brasileira de cartografia Vol. 57; no. 3; pp. 168 - 179
• Main Authors: João F. C. da Silva, Aluir P. Dal Poz, Ricardo L. Barbosa, Rodrigo B. de A. Gallis
• Format: Journal Article
• Language: English
• Published: Universidade Federal de Uberlândia 01.12.2005
• Subjects: aerial images; Dynamic Programming; feature extraction; roads
• ISSN: 0560-4613; 1808-0936

This article proposes a novel road extraction methodology from digital aerial images. The innovation is based on the dynamic programming (DP) algorithm to carry out the optimization process in the object space, instead of doing it in the image space such as the DP traditional methodologies. The basic principle is inspired on the photogrammetric plotter functioning. In both cases, the features are traced in the object space, which implies that a rigorous mathematical model is necessary to establish a relationship between image and object space points. However, the proposed methodology does not require the operator to compile the features in the object space. It just requires that the operator measures a few seed points in the image space to describe sparsely and coarsely the roads, which must be transformed into the object space to make possible the initialization of the dynamic programming optimization process. Flexibility is one of the most important advantages of the proposed methodology, that can be operated in different modes (mono-plotting or stereo-plotting), and with several image types, including multisensor images. This paper presents the basics of the novel methodology for road extraction in the object space, as well as the details of the semi-automatic road extraction process using a single image, along with the experimental results.