SYNTHETIC APERTURE RADAR IMAGE FORMATION SIMULATOR FOR

EDUCATIONAL PURPOSE
Musyarofah1, Rahmat Arief1, Muchammad Soleh1
1
Remote Sensing Technology and Data Center, Indonesian National Institute of Aeronautics and
Space (LAPAN),
Jl. Lapan No. 70 Pekayon, Pasar Rebo, Jakarta Timur 13710, Indonesia
email: musyarofah@lapan.go.id; rahmat.arief@lapan.go.id
ABSTRACT
Design and development of SAR system needs an extensive understanding of SAR technology, it
takes time and expensive cost. One of prior challenge in SAR technology is process of twodimensional SAR image formation from raw data SAR. SAR sensor generates huge volume of raw
data that should be processed to make two-dimensional image using spesific SAR image algorithm.
This paper introduces SAR image formation simulator that was developed to simplify SAR system
parameter learning, data processing and SAR image formation analysis for researchers, engineers and
students. This simulation is implemented by using MATLAB software. It was developed modularly,
comprises of three moduls, (1) preparation of Radarsat-1 (SAR) raw data and formation of SAR raw
data from point target (2) implementation of image formation algorithm, such as RDA and (3)
evaluation of SAR image quality. This simulation also can be used for further analysis and SAR image
formation algorithm comparation.
Keywords: SAR, image formation algorithm, simulator
INTRODUCTION
Recently, remote sensing techique using synthetic aperture radar (SAR) has became more popular
and has been widely utilized. Understanding of SAR technology is needed as an approach to develope
SAR system and SAR raw data processing. The basics of SAR raw data processing are very important
and have to be understood, to generate focused SAR image for further different application. Focused
SAR image is generated by using SAR raw data processing, namely SAR simulator, that implements
SAR image formation algorithm. There are some image formation algorithm that have been developed
and used to effectively process SAR data from its raw signal into well focused images, such as Range
Doppler Algorithm, Chirp Scaling Algorithm, Omega-K Algorithm, and SPECAN Algorithm.
This paper describes SAR image formation simulator that developed in Remote Sensing
Technology and Data Center, LAPAN. It is written to ease the beginners to understand the basics of
SAR system, SAR image formation algorithm and its simulation. It also gives an understanding of
SAR image quality and its parameters. The algorithm used in this simulation is Range Doppler
Algorithm (RDA) that implemented in simulator using MATLAB source code. The raw data used are
Radarsat-1 and simulated point target.
SAR RAW DATA PROCESSING
Before getting into SAR raw data processing, it is needed to figure out SAR data acquisition
system. It helps to discern how the data is obtained from a point target surface.
Radar is an active system that transmits electromagnetic radiation and receives the backscattered
radiation from the illuminated point target surface. Basically, there are three operating modes of SAR
system, that is stripmap, scan and spotlight SAR modes. Operating mode that defined in this paper is
stripmap SAR mode. It is the most popular operating mode of SAR system. In this mode, the radar
antenna point along fixed position direction according to the flight platform path, and the antenna
footprint covers a strip on the illuminated point target surface as platform moves (see Figure 1).
In context of SAR geometry system, the axis directions are defined as azimuth direction, which is
aligned with the relative platform velocity vector (satellite flight direction), and range direction, that is
a distance from the radar antenna to the target which perpendicular to the azimuth direction (see
Figure 1). In the SAR system, antenna transmits pulses at the certain pulse repetition frequency (PRF)
to the point target surface. When a point target is illuminated by SAR system, it reflects backscattered
energy that spreads in the range and azimuth direction. The backscattered energy from point target is
in the form of electromagnetic signal, it is received by antenna receiver and saved in the SAR system.

Figure 1. Geometry of SAR acquisition system

The objective of SAR raw data processing is to compress the signal into a single pixel to
reconstruct the point target. The signal received in SAR system contains amplitude and phase
information of point target backscatered that will be tranformed into an image using SAR signal
processing algorithm. As mentioned above, SAR signal processing algorithm used in this paper is
RDA, which is the algorithm that designed to achieve block processing eficiency, using frequency
domain operations in both range and azimuth, while maintaining the simplicity of one-dimensional
operations. It takes advantage of the approximate separability of processing in these two directions,
allowed by the large differences in time scales of the range and azimuth data, and by the use of range
cell migration correction (RCMC) between the two one-dimensional operations. The steps consist of
range compression, RCMC and azimuth compression (see Figure 2).

Figure 2. Range Doppler Algorithm Flowchart

For SAR imaging system, the pulse transmitted is a long duration linear FM pulse called chirp.
And the returned signal is convolved with the chirp. To recover the desired signal from the target, it is
needed to deconvolve the chirp through matched filter. Matched filter provides different time delay for
each frequency component and at the output, all the frequency component arrives at the same time to

compress the energy of signal into a narrow pulse, this process is called pulse compression. When a
point target is illuminated by SAR system in azimuth direction, all the reflected signals from that
target must be situated in the same range bin in different echoes, but practically it does not happen due
to the variation of the range between the radar and the scatterer, called Range Cell Migration (RCM).
To compress signal energy in azimuth direction, we need to align signal energy within a single range
bin, where range bin depends on the range at which point target is in the centre of antenna beam. This
conducted by process called Range Cell Migration Correction (RCMC).
In RDA, the received echoes (signals) in range direction are brought into frequency domain using
Fast Fourier Transform (FFT) and then multiplied with the range matched filter (range reference
function) for range compression. The range compressed data is then brought back to range Doppler
domain (range time and azimuth frequency domain) using azimuth FFT. The final step is azimuth
compression, it is conducted by multiplying the azimuth reference function in frequency domain with
range compressed data and then azimuth IFFT is done to transform the data back into time domain.
Finally, the image is generated where each pixel represents the energy from a single point on the
ground surface.
SAR IMAGE QUALITY PARAMETERS
The main objective of the SAR image quality analysis method is to provide a tool for study of the
focusing performance of different processing algorithms and the influence of key parameters in the
quality of the final image. SAR image quality analysis method comprises of SAR focusing
performance and radiometric quality of the image. However, in this paper we limit the discussion only
on SAR focusing performance.
SAR image quality parameters that related to SAR focusing performance involve spatial
resolution, peak side lobe ratio (PSLR) and integrated sidelobe ratio (ISLR). SAR image quality refers
both to impulse response properties of a radar/processor combination, and to the response of the
system to distributed scatterers. SAR response to a point target, assuming negligible background
reflectivity and thermal noise, is commonly referred to impulse response function (IRF). The analysis
of the signature of a point target in a SAR image allows the determination of several parameters that
related to SAR spatial resolution and the presence of side lobe peaks. The IRF is a sinc function with a
mainlobe and many secondary lobes. It is two-dimensional signal appearing in a processed image as a
result of the compression of returned energy from a point target. Spatial resolution of SAR image is
defined as impulse response width (IRW), refers to the width of the impulse response function at a
power level 3 dB below the peak function (see Figure 4).

Figure 3. Representation of IRF of a point target

A side lobe of the IRF is any local maximum other than those within the contour around the peak,
which passes through points 3 dB below the main lobe peak. Side lobes are measured relative to the
main lobe peak. The PSLR is defined to be the ratio of the maximum sidelobe level and the main lobe
level. The PSLR is calculated as
= 10 log

(1)

where stands for the peak intensity of the most intense sidelobe and
stands for the peak
intensity of the main lobe.
The ISLR is defined to be the ration of integrated energy in the side lobes region of the two
dimensional (range and azimuth) IRF relative to the integrated energy in the main lobe region. The
ISLR can be obtained by integrating the power (magnitude squared) of the impulse response over
suitable regions of sidelobe. If the main lobe power is
and the total power
, the onedimensional ISLR is then
= 10 log

!
!

" (2)

where the numerator is the total power of the sidelobes.

RESULTS AND DISCUSSION
The digital SAR processor is a computer program that converts the raw signal data into singlelook complex (SLC) image. The step in the simulation of digital signal processing for SAR image
formation using MATLAB source code that conducted is described in Figure 5.
Parameter
file

Raw signal
data file

2-Dimensional Fast Fourier

Transform

Range
Compression

Range Inverse Fast

Fourier Transform

Range Cell Migration

Correction

Azimuth
Compression

Azimuth Inverse Fast

Fourier Transform

Single Look
Complex (SLC)
image data file

Figure 4. Overview of RDA simulation of Radarsat-1 raw data processing using MATLAB
The raw data obtained by SAR system is not an image in any form, it is a data file consists of
signal properties of transmitted and received signal, plus auxiliary data defining parameters regarding

the satellite orbit, speed and other characteristics. The first step of SAR raw data processing is to read
the satellite raw data file and extract desired data. Radarsat-1 raw data is stored in the CEOS format.
The first 192 lines of the file are file header which contains information regarding time and date of
data acquired, and also satellite orbit information. Then, there are 50 lines where auxiliary data is
stored. After these lines, the main data body located, where transmitted and reflected data is stored.
Radarsat-1 data is stored on 8-line blocks structure, it means that received signal from 8 transmissions
in azimuth direction is stored together with one replica of transmitted signal. The collected data is
encoded and then stored in the raw data file.
To extract a set of desired data, relevant lines of data file are read, then the real and imaginary
parts of the complex signal is extracted, the data is decoded and then, extracted data is saved in the
shape of a matrix consisting of complex numbers. In this program if we want to extract a certain
window of the whole data, the program first finds the starting line of data file corresponding to the
selected window. The starting line has to be adjusted, since data is stored on the basis on 8-line blocks.
Therefore the selected window automatically moves to start line of the corresponding data block. Then
program starts extracting data from that point. The reason is the size of source data file that very large
about 400 MB, which contains data for about 180 million cells, means 360 million real and imaginary
numbers. It is almost impossible to load the whole data on a PC and then look for the selected
window.
When the staring point of extraction is defined, program starts reading rows of the selected
window completely, at the same time it separates the replica data and radar echo data. Then the
program extracts exact selected window by cutting out unnecessary extracted columns and then
demodulates the data to complex numbers containing I and Q channels of the reflected signal.
After reading raw data file and importing parameter file, next step is range compression. It is a
recovering a sharp radar pulse by deconvolution of the chirp. Range compression is performed with a
fast convolution when the data are in the azimuth time domain. In other words, range FFT is
performed followed by range matched filter multiply, and finally a range IFFT to complete range
compression. The next step is range migration using FFT. Range migration, also called RCMC,
straightens out the trajectories so that now it run parallel to azimuth frequency axis. And the last step is
azimuth compression, which is to focus the data in azimuth by accounting for the phase shift of the
target as it moves through the aperture. It is conducted by generating a second frequency-modulated
chirp where the chirp parameters depend on the velocity of the spacecraft, the PRF and the absolute
range. The chirp is fourier transformed into Doppler space and multiplied by each column of rangemigrated data. The product is inverse fourier transformed to provide the focused image. The display
results of each step of Radarsat-1 image formation using RDA are shown in Figure 5.

Figure 5. The display results of RDA simulation of Radarsat-1 raw data processing using
MATLAB
To understand SAR image formation is not only able to be conducted by simulation using
Radarsat-1 raw data. It also can be conducted by simulation using simulated point target. A simulated
point target is a point target that generated with source code in MATLAB by using SAR system
parameters. Therefore, this simulated point target approaches the real point target in ideal isolated
condition. The point targets are arrayed in a Cartesian type coordinate system space defined by range,
azimuth, and altitude as analogs of x, y and z directions. The altitude direction is omitted in the two-

dimensional simulation. The platform in this simulation is an antenna attached to a plane traveling at
an orbital velocity, Vs, along the azimuth direction and at the midpoint in the flight, the distance to the
target equals the range of closest approach or minimum range to target, RO in the simulation and
denoted by XC, see Figure 6. The curvature of the earth is considered negligible and the orbital
velocity is approximately equal to the platform velocity, Vr. The velocity of the radar beam along the
ground is Vg. The operating mode used in this simulation is stripmap SAR mode, means the radar dish
is pointed in a fixed direction during the duration of the flight (see Figure 6). As the platform traverses
the azimuth, the radar beam sweeps along the ground. The part of the radar beam touching the ground,
shown in the circles to the right, is calles the beam footprint. The beamwidth is the cone extending
from the antenna to beam footprint. Targets in the beam footprint reflect back radar signals which are
then received by the antenna. The obtained radar reflections are processed with the RDA, to obtain the
final SAR image.

Figure 6. Stripmap SAR geometry that used in simulated point target formation
The display results of the image formation simulation with simulated point target resolution is
shown in Figure 7.

Figure 7. The display results of simulated point target raw data image formation using RDA

In this simulation, we also measured SAR image quality parameters of the both raw data,
Radarsat-1 raw data and simulated point target raw data. The spatial resolution, PSLR, ISLR and peak
intensity are measured in this simulation. All measurement and simulation are conducted by using
MATLAB source code. First, we derived IRF graphical representation of SAR image resulted in the
previous simulation (See Figure 8). IRF graphical representation shows impulse response of echo
signal of point target that received by SAR antenna. Higher peak level of IRF main lobe, stronger
signal received, and vice versa.

Figure 8. IRF graphical representation of Radarsat-1 point target

Figure 9. IRF graphical representation of simulated point target

From IRF graphical representation, the value of each SAR image quality parameters is derived and
measured. The value results are shown in Table 1.
Table 1. The value of SAR image quality parameters resulted in the simulation
3db Resolution
[pixel]

PSLR [dB]

ISLR [dB]

Peak [Intensity]

Radarsat-1 Point
Target Raw Data

1.932

-8.873

-14.964

310.984

Simulated Point
Target Raw Data

1.979

-9.423

-15.626

1307.794

Data type

The value resulted is qualified enough and the image resulted is quite focused to be used in further
SAR data application. The modul package of this simulator should be more developed and completed

with graphical user interface (GUI), hence it will be more user-friendly, and eases the beginners to
study and understand the algorithm and the steps of SAR image formation.
CONCLUSION
SAR simulation using RDA that implemented in MATLAB source code is easy for the beginners
to understand the algorithm and the steps of SAR image formation. SAR image formation using RDA
results quite focused image. The measurement of SAR image quality parameters is also conducted in
this simulation using MATLAB source code. The image resulted is qualified enough to be used for
further application. It is proven regarding to the value of its quality parameters shown in Table 1. For
further development, the modul package should be completed with GUI and additional modul such
windowing to increase SAR image quality. The other algorithms should be added and implemented in
the modul package, so RDA also can be compared with the other algorithms.
REFERENCES
[1] Cumming, I., & Wong, F., 2005, Digital Processing of Synthetic Aperture Radar Data:
Algorithms and Implementation, Artech House.
[2] Lipscomb, J., 2002, SAR Image Quality, Alaska.
[3] Skolnik, M., 2001, Introduction to radar systems, 3rd edition, McGraw-Hill.
[4] Zenere, L. M.,2012, SAR Image Quality Assessment, Cordoba, Universidad Nacional de
Cordoba, Spain.
[5] Curlander, J. C. and McDonough, R. N., 1991, Synthetic Aperture Radar: Systems and Signal
Processing, John Wiley & Sons, New York.
[6] Deo, R., et.al., 2012, MATLAB Based SAR Signal Processor for Educational Use, IGARSS
2012, IEEE 978-1-4673-1159-5/12/$31.00 (page 5318-5321).
[7] Jost, R. J., 2005, MATLAB-Based Toolkit For An Introductory Course in SAR Image Processing,
IEEE 0-7803-8882-8/05/$20.00
[8] Akliouat, H., et.al., 2007, Synthetic Aperture Radar Image Formation Process: Application to a
Region of North Algeria, Proceeding of Envisat Symposium 2007, Montreaux, Switzerland 23-27
April 2007.
[9] Datsgir, N., 2007, Processing SAR Data Using Range Doppler and Chirp Scaling Algorithms,
Master Thesis of School of Architecture and Built Environment, Royal Institute of Technology
(KTH), Stockholm, Sweden.
[10] Sandwell, D. T., 2002, SAR Image Formation: ERS SAR Processor Coded in MATLAB.
[11] Schultz, M., 2009, Synthetic Aperture Radar Imaging Simulated in MATLAB, Thesis of Faculty
of the California Polytechnic State University, San Luis Obispo.