Learn why you were pulled over, quantify the stealthiness of your favorite model aircraft, or see what various household items look like at 10 GHz. In this post we will describe the basics of Synthetic Aperture Radar (SAR) imaging, beginning with a historical perspective, showing the state of the art, and describing what can be done in your garage laboratory. Lets image with microwaves!
The History of SAR
Ground mapping (or imaging the ground terrain) using microwave radar was done routinely in the Second World War by the Royal Air Force for the purpose of navigation and bomb laying using the H2S radar system. The H2S used a large aperture rotating antenna in the belly of a bomber aircraft. This antenna would rotate in circles with its beam directed toward the ground. Range to target was plotted in a plan position indicator (PPI, or a radar screen as most would recognize it) showing what was below and around the aircraft.
The angular resolution of this radar set depends on the antenna aperture size (e.g. antenna size). The bigger the aperture the finer the angular resolution, just like the reflector on a flashlight provides a tighter light beam the larger it is (this is why spotlights shine tight beams well into the sky). A typical example of an H2S radar image is shown below recorded during s bombing raid over Berlin. In this image the river is clearly visible as well as other blob-like targets which are landmarks that a trained operator would recognize.
Earlier versions of the H2S were at S-band (3 GHz) and later higher resolution sets were at 10 and 24 GHz (for an interesting read on this technology, Echos of War: The Story of H2S Radar).
Synthetic Aperture Radar (SAR) is a modern ground mapping technique where high resolution is achieved by a very large aperture that is synthesized over the flight path of an aircraft. This is done by recording reflected radar pulses at known locations along the flight path. The radar must accurately know the aircraft’s position and back-out perturbations in flight path so that all scattered pulses are aligned in time and phase. After this a SAR imaging algorithm is applied to the data to process an image.
Developments in SAR Technology
This technique was first developed in 1957 using photographic film to record the radar data and an image processor made from lenses. Today digitizers and other data acquisition equipment can store data for offline processing or even process imagery in real-time.
State of art airborne SARs include the MIT Lincoln Laboratory LIMIT system (PDF), which operates at X-band (10 GHz) and is mounted on an old 707 aircraft for testing advanced SAR imaging concepts.
Another is the Sandia National Laboratory’s Ka Band SAR imaging system (to see an amazing portfolio of airborne SAR imagery visit here PDF), an example airborne SAR image from this system is shown below.
SAR imagery appears to be nearly photographic but it is not a photograph, it is a 2D hologram. Unlike a satellite image the radar is not measuring the target scene from above it is measuring from the side at a fairly significant distance. The resulting image is a birds-eye view with many shadows where each pixel is mapped directly to the aircraft’s flight path in range and cross-range.
Most recently, small and light weight airborne SAR imaging systems weighing only a few lbs have been developed for micro-UAVs, for example the NanoSAR imaging system manufactured by IMSAR.
Create your own SAR imaging system.
Airborne SAR imaging is beyond the means of most hackers and hobbyists. The good news is that you can do it yourself with better resolution if you limit the scope of the problem and reduce maximum range, power, and the complexity of your radar sensor. To achieve this consider the rail SAR imaging system. In this, an ultrawideband (UWB) radar device is mounted on a long linear stage (typically 6′ to 8′ in length). The radar pulses once, moves, pulses again, each echo is recorded. This process repeats itself along the rail until a complete data set is acquired.
For the UWB radar sensor you can use one of the sensors described in my previous post that is either an impulse or an FMCW radar or create your own. For the linear rail stage you can use anything from a Genie garage door opener assembly (which contains a lead screw inside of a long aluminum extrusion with a car that rides on the threads) to one stage on a full-size CNC router table.
Make your own from junk parts
One example of a hacked-together rail SAR is the ‘backyard SAR’ imaging system, where an X-band UWB FMCW radar front end was mounted to an 8′ long linear stage built from a Genie garage door opener, a cordless drill transmission, and a stepper motor following the block diagram shown. X-band microwave components were acquired at hamfests.
To process data from a rail SAR like this follow the procedure outlined in the Range Migration Algorithm chapter from Spotlight Synthetic Aperture Radar: Signal Processing Algorithms, which follows these steps:
- Cross range discrete Fourier transform (DFT).
- Apply matched filter.
- Perform Stolt interpolation.
- 2D IDFT into image domain.
When implemented correctly this will result in the imagery shown below, achieving approximately 1×1” resolution at X-band with approximately 5 GHz of chirp bandwidth.
Build the coffee can radar kit
To make SAR imaging accessible the MIT ‘coffee can’ radar course was developed, where you can SAR image with the coffee can radar. The goal of the SAR imaging experiment was to show students it is possible to differentiate in both rang and cross range when imaging some very large targets.
The coffee can radar does not produce the best imagery but it shows a concept to students. To acquire an image, it is placed on a linear track with a tape measurer for a position reference. This could be a length of 2×6” or a straight rail somewhere. The radar is manually moved in 2” increments where a toggle switch on the side mutes the synchronization signal output, showing the computer that the radar has moved.
Resulting in imagery comparable to that shown below.
Give it a try, but be sure to image a large target scene. The algorithm is already written and the procedure is straight forward (scroll down to ‘Experiment 3: SAR imaging’).
It is not trivial to design, build, and write a an imaging algorithm for your backyard rail SAR. Caveats to implementation and processing include having to scale to your wavelength range, the need for calibration to a point target (a large pole or similar), use of coherent background subtraction, and other processing techniques. But we can philosophize about these all day, the best way to learn is to try it yourself:
- Learn by doing, build the MIT Coffee Can Radar and try the SAR imaging experiment.
- For a quick-read technical background read Chapter 4 and for details on numerous practical examples Chapter 5 in the book Small and Short-Range Radar Systems (use promo code EEE24 for discount).
- Process a SAR image right now. Download data sets for X and S-band and their associated processing algorithms written in MATLAB. With this you will learn how to apply calibration and coherent background subtraction.
- Need help? Post your questions to the Tin Can Radar Forum.
With these resources, patience, perseverance, and coffee anyone can create a SAR imaging system in their garage.
Gregory L. Charvat, is author of Small and Short-Range Radar systems, co-founder of Butterfly Network Inc., visiting research scientist at the Camera Culture Group MIT Media Lab, and editor of the Gregory L. Charvat Series on Practical Approaches to Electrical Engineering. He was a technical staff member at MIT Lincoln Laboratory from September 2007 to November 2011, where his work on through-wall radar won best paper at the 2010 MSS Tri-Services Radar Symposium and is an MIT Office of the Provost 2011 research highlight. He has taught short radar courses at the Massachusetts Institute of Technology, where his Build a Small Radar Sensor course was the top-ranked MIT professional education course in 2011 and has become widely adopted by other universities, laboratories, and private organizations. He has developed numerous rail SAR imaging sensors, phased array radar systems, and impulse radar systems; holds several patents; and has developed many other radar sensors and radio and audio equipment. He earned a Ph.D in electrical engineering in 2007, MSEE in 2003, and BSEE in 2002 from Michigan State University, and is a senior member of the IEEE, where he served on the steering committee for the 2010 and 2013 IEEE International Symposium on Phased Array Systems and Technology and chaired the IEEE Antennas and Propagation Society Boston Chapter from 2010-2011.