Share Email Print
cover

Proceedings Paper • new

Automatic change detection using very high-resolution SAR images and prior knowledge about the scene
Author(s): C. Villamil Lopez; T. Kempf; R. Speck; H. Anglberger; U. Stilla
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Change detection using very high resolution SAR images is an important source of information for reconnaissance applications. Modern SAR sensors are capable of acquiring many images in short periods of time, which creates the need for a reliable automatic change detection method. In this paper, we will describe a new automatic change detection approach that combines very high resolution SAR images with prior knowledge about the imaged scene. In this case, the prior knowledge about the scene will come from vector maps, which can be obtained from a Geographic Information System (GIS). These vector maps will allow us to determine which regions are of interest for the change detection, and what kind of changes/objects can be expected there. The algorithm described in this paper will be applied to a time series of high resolution TerraSAR-X images of a port with military shipyards, and used to automatically detect ship activity and extract information about the detected ships. In this case, the vector maps were obtained from a Geographic Information System (GIS) containing map data from OpenStreetMap

Paper Details

Date Published: 1 May 2017
PDF: 9 pages
Proc. SPIE 10188, Radar Sensor Technology XXI, 1018805 (1 May 2017); doi: 10.1117/12.2262529
Show Author Affiliations
C. Villamil Lopez, Deutsches Zentrum für Luft- und Raumfahrt e.V. (Germany)
T. Kempf, Deutsches Zentrum für Luft- und Raumfahrt e.V. (Germany)
R. Speck, Deutsches Zentrum für Luft- und Raumfahrt e.V. (Germany)
H. Anglberger, Deutsches Zentrum für Luft- und Raumfahrt e.V. (Germany)
U. Stilla, Technische Univ. München (Germany)


Published in SPIE Proceedings Vol. 10188:
Radar Sensor Technology XXI
Kenneth I. Ranney; Armin Doerry, Editor(s)

© SPIE. Terms of Use
Back to Top