Examples of previous simulation systems are by Holtzman et 
al., (1978) and Kaupp et al. (1982). The simulation discussed 
here emphasizes the correct image geometry and uses homogenous 
backscatter curves to model the radiometry as a funetion of the 
incidence angle. 
Radar Image Simulation 
The task of image simulation consists of two separate tasks -- a 
geometric one vs. a radiometric one. The geometric part of the 
simulation relates the object and image point. The radiometric 
task assigns grey values to each image pixel according to the 
properties of the corresponding object cell and an additional 
backscatter curve. 
(a) Geometric Imaging Model 
Two types of algorithms may be applied to establish the 
relationship between object and image point addresses. The 
straight forward approach is to start in the object space (the 
digital elevation model, DEM) and map the object space 
coordinates (DEM grid points) into the image plane by applying 
the radar equation 
2 
e? itgenégy aua 
r ... slant range 
d ... distance between nadir and point to be imaged 
B ... flight altitude 
h ... height of imaged point 
This results in non-equidistant grid points in the image as image 
space coordinates. Some sort of interpolation has to be applied 
to create a regularly spaced output. 
The other approach to image simulation would be an image space 
algorithm as opposed to the object space algorithm described 
above. One starts with the equidistant image space coordinates 
(x,y) for the output image. These need to be converted to the 
imaging time t and slant range r. Time t serves to derive the 
platform position S and corresponding velocity vector y from 
given flight recordings. 
The geometry created by the platform position S, the slant range 
r and known squint angle T is the radar projection circle at the