Abstract
Digital terrain models (DTM) are an important source for the detection
of areas which are susceptible for natural hazards. In order to exploit
the potential of stereo satellite imagery for time series analysis,
addressing registration and elevation errors is a crucial topic. For our
analysis we used different spaceborne multitemporal stereoscopic images
- Corona (1962 and 1970) and Cartosat-1 (2007). The Cartosat-1 DTM
showed the highest absolute accuracy (less than 20 m), whereas the older
Corona DTM exposed the highest RMSz error. Due to differing acquisition
techniques DTMs of the same area will never be identical. Therefore, we
propose a relative vertical accuracy assessment with the Cartosat-1 DTM
as master DTM. A comparison between the reference and input DTMs
revealed inaccuracies, which were addressed by convenient statistical
parameters and adjusted by means of trend surfaces. To eliminate
outliers the one and a half time interquartile range was calculated.
Original language | English |
---|---|
Title of host publication | Proceedings of ESA Living Planet Symposium |
Pages | 195 |
Volume | 686 |
Publication status | Published - 1 Dec 2010 |