Simulation of 3 - d terrain - an approach to evaluate SAR interferometry technique for DEM
M. Sudhakar Rao, Satyendra K. Khare and K. S. Rao
Centre of Studies in Resources Engineering, IIT
Mumbai - 400 076
SAR Interferometry is a new technique being used to generate DEM. SAR system records both amplitude and phase of the back scattered signals, from which Phase and amplitude images are produced. Single SAR Image phase value doesn't have any practical value. Two SAR Images of the same area taken from the same satellite or from different satellites have different phase values. The phase values of two SAR images have the application in generating DEM by taking phase differences of the corresponding pixels on the two images. The phase difference is because of change in the Elevation of the terrain and change in the satellite trajectory. The Technique includes, preparation of SLC image from the raw data set, Registration of the slave image to the master image, Generation of the interferogram from the phase differences of the two images, Phase unwrapping, selection of the Ground Control Points (GCP) to generate the DEM from the unwrapped image and Geo-coding of the DEM.
The present study is carried out to understand the above-mentioned technique. A dome type terrain with 300-m elevation and spread of 8 Km on 10 ? 10-Km area was considered. Satellites at a height of 780 Km and 780.1 Km and a horizontal shift of 200 m were considered with a look angle of 230. Phase values were calculated for different points on the terrain at a slant range resolution of 7.5 m. Thus an image of 500 ? 500 pixels was generated for each satellite track. The phase differences were calculated at each pixel for both the satellite positions. This forms the interferogram image. For base line of 223 m, 3.5 fringes were developed. Each fringe corresponding to about 86 m elevation on the terrain. Phase unwrapping was carried out at every pixel and an unwrapped image file is created in which no further fringes are seen. The unwrapped image and the height of ambiguity are used to convert the unwrapped phase values to height values. Height of ambiguity depends on the baseline between the two satellite orbits. The heights thus generated are then geocoded to get the actual position on the terrain at a ground resolution of 20 m. The height values calculated from this exercise are in good agreement with the actual values with in an error of 0.5 m.The details of simulation, computation of errors and Geo-coding will be given in the full paper.
Page 1 of 1