Share Email Print
cover

Proceedings Paper

Extraction of linear features in SAR images for geographical map updating in a tropical forest context
Author(s): V. P. Onana; Emmanuele Trouve; G. Mauris; Jean-Paul Rudant; J. Mvogo Ngono; Emmanuel Tonye
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In this paper, a new linear features (LF) extraction method is proposed to update old geographical maps using Synthetic Aperture Radar (SAR) data. This method is dedicated to the context of tropical rain forest where LF such as roads or railways often appear as long segments poorly contrasted and partially hidden in a homogeneous environment. Most classical SAR road extractors operate locally in the spatial domain using a point to point approach that does not take into account the “a priori information” provided by old maps. Moreover, often developed in a different context, they are not well adapted to the extraction of long segments which are most of the time difficult to follow, even for trained human operators. To overcome these problems, the proposed method is based on the Localized Radon Transform (LRT) which performs limited boundary geometrical integrals along straight lines. In the transformed domain, LF have a specific signature: they appear as strongly contrasted structures, easier to extract with a constant false alarm rate operator: the ratio of local mean. The “a priori information” is integrated by two parameters issued from the old map: the approximate direction angle and the minimum length of the LF. Experimental results show the robustness of this method with respect to the poor radiometric contrast and hidden parts. According to these results, this method can be proposed as a complementary tool to detect LF in SAR images.

Paper Details

Date Published: 21 December 2000
PDF: 11 pages
Proc. SPIE 4173, SAR Image Analysis, Modeling, and Techniques III, (21 December 2000); doi: 10.1117/12.410646
Show Author Affiliations
V. P. Onana, Univ. de Marne-La-Vallee, Ecole Polytechnique de Yaounde and Univ. de Savoie (France)
Emmanuele Trouve, Univ. de Savoie (France)
G. Mauris, Univ. de Savoie (France)
Jean-Paul Rudant, Univ. de Marne-La-Vallee (France)
J. Mvogo Ngono, Univ. de Marne-La-Vallee (France)
Emmanuel Tonye, Ecole Polytechnique de Yaounde (Cameroon)


Published in SPIE Proceedings Vol. 4173:
SAR Image Analysis, Modeling, and Techniques III
Francesco Posa; Luciano Guerriero, Editor(s)

© SPIE. Terms of Use
Back to Top