Satellite Radar Interferometry for DEM Generation

By Amy Williams

Supervised by Dr Linlin Ge

Edited by J. M. Rüeger

October 2003


Topographic mapping traditionally has been created by using stereo pairs of photographs taken from aircraft and satellites. Unfortunately, these optical systems do not work if the earth’s surface is covered by cloud. Radar, however, is able to penetrate the cloud cover and to map areas, that could not be mapped previously (e.g. some tropical areas) with a high accuracy and resolution. Radar allows data to be acquired at any time, day or night, regardless of weather conditions.

A digital elevation model (DEM) can be defined as a set of points stored in a digital format, that represent the earth’s surface in three-dimensional coordinates. DEMs are essential in creating topographic data. Topographic data allows scientists to study earth movements, lava and ice flows, flooding and drainage. Also, this data is essential for military and emergency services for training and real-time operations.

Figure 1: Georeferenced DEM using image pair 22434_2761

Background on Radar

RADAR stands for RAdio Detection And Ranging. It is an active remote sensing system, as it provides its own source of radiation. It uses microwave electromagnetic radiation, emitted by an antenna, to illuminate the earth’s surface. It then measures the backscatter (radiation reflected back) from the surface. The distance to the target is computed by measuring the time taken for the radar signal to travel to the target and back. Radar remote sensing uses wavelengths from the microwave part of the spectrum, namely wavelengths from 1 mm to 1 m.

Synthetic Aperture Radar Interferometry

Radar interferometry (or Interferometric Synthetic Aperture Radar, INSAR) is a technique, which relies on the measurement of the phase of the backscattered signal. By determining the difference in the phase of the return signals from two different antennas, one can determine topographic heights and sea surface velocities. The signals transmitted from one antenna are simultaneously received by both antennas.

Project Description

The main aim of this thesis was to generate a high accuracy and high-resolution digital elevation model (DEM) using satellite radar interferometry from ERS tandem data. The DEMs generated were then tested and validated against known ground survey data. The test area was a mining area to the Northwest of Wollongong. There are two mines in the area, namely the Appin Colliery and the Westcliff Colliery. The survey data, that was used for comparison, are subsidence lines monitored by the mines.

Three different methods of generating a DEM were compared. The optimal method, for achieving accurate results, was determined. It was hoped to achieve a high accuracy DEM, that can be used directly to monitor ground subsidence above underground mines and tunnels by using satellite radar interferometric data.

Figure 2: Aerial photograph of the test site.


A program developed by Canadian Company Atlantis, called Ev-INSAR, was used to generate the DEMs. The satellite data was entered into the program and several porting steps were carried out, namely the coregistration of the images, the generation of the interferogram, the phase unwrapping and the generation of the DEM. The product was then entered into a GIS for georeferencing and finding the heights along the survey lines. These heights were then compared against ground survey data using MS Excel.


Since the best results, that were achieved, were only to the order of 4-6 m, and since subsidence needs to be established at centimetre level, it is my opinion that this method of DEM generation is not practical or useful for the direct monitoring of ground subsidence above mining and tunnel areas. However, this method may be very useful in the generation of topographic maps and to create a more accurate DEM for the whole of Australia (replacing the current AUSLIG 9" DEM).

Of the three different types of DEMs produced in the project (DEM without reference DEM, DEM using a 1" reference DEM and DEM using a Tandem DEM as the reference), the best results came from the DEM without the use of a reference DEM. This was unexpected as logic would predict that a reference DEM would generate more accurate results.

Further Information

For more information, please contact:

Dr Linlin Ge

School of Surveying and Spatial Information Systems
University of New South Wales

Phone: +61-2-9385-4188
Fax: +61-2- 9313-7493

Miss A. J. Williams