Depth Map Estimation from Flight Image Series using Multi-View Along-Track Stereo

Conference paper


Ioana Gheta
Christian Frese
Wolfgang Krüger
Günter Saur
Norbert Heinze
Michael Heizmann
Jürgen Beyerer


A. Grün, H. Kahmen (eds.), Optical 3-D Measurement Techniques, Zürich, vol. 2, 2007.




8th Conference on Optical 3-D Measurement Techniques, Zürich, July 9 - 12, 2007

3-D reconstruction of landscape or cultural sites is a frequent task. One of the most common ways to solve the problem is to fly a camera over the area of interest and fuse the obtained image series. Practically, there are several challenges to overcome due to the unknown camera and flight parameters that influence the process. This contribution proposes a way to overcome these difficulties by applying a multi-view algorithm for estimating depth. Other than the usually employed stereo based algorithm, the proposed method uses more than two images for estimating depth. In such a way, the result becomes more reliable and robust. The performance of the method is shown on image series gathered by simulating flight over a model landscape.