A SOFTWARE upgrade to existing radar systems will allow aircraft to map landscapes and buildings in 3D, and make it easier to spot part-hidden objects on the ground.

The software augments synthetic aperture radar (SAR), a technique that has been around since the 1950s. Airborne SAR works by sending radar signals towards each spot on the ground from several different points along a plane’s flight path. Signals bouncing back from a particular object on the ground are combined to create a 2D image as if it had been taken from a single viewpoint – and in more detail than would be possible from a single source.

In the new approach, being developed in a project called Exploitation of Geometric Diversity, the aircraft flies in a curved or circular path rather than a straight line. This provides a series of images of the same object on the ground from a number of slightly different viewpoints, providing detailed depth information about the object. Software then produces a 3D image from the returning signals.

This is similar to the way stereoscopic photography and 3D movies use two images of the same thing taken at different angles to give the appearance of depth. Each pass only provides a small amount of data, but this can be built up over multiple passes.

Details of the project appear in US air force contracts awarded to two US companies under the Small Business Innovation Research programme.

“If enough fly-bys are collected, with lots of arcs, full tomographic imaging can be carried out, akin to medical imaging,” says Keith Morrison, who investigates radar techniques at Cranfield University in the UK. “You can create an image slice-by-slice in height, and run it as a movie just like a medical image through a brain scan.”

This could be used, for example, to locate a crashed aircraft hidden beneath a forest canopy. Some 2D radar systems are able to penetrate foliage, but they often produce a noisy image with a lot of “clutter”, as signals from the trees and objects on the ground are combined in a single image. The new 3D radar’s software could “strip off” the forest canopy, to get a much clearer image of any object on the ground.

3D radar could also be used to map the insides of buildings – so long as the building does not have a metal roof that would reflect the radar signal – or measure the thickness of ice sheets or oil slicks.

The curved flight path also greatly improves the chances of picking up “glints” – strong radar reflections from certain angles that may reveal the location of objects that would otherwise be invisible. Initially, the 3D radar will be carried on Predator drones. Trials are expected to be completed by 2013.

To convert conventional SAR radar reflections into a single image, Predators already carry signal-processing hardware with computing power equivalent to about 100 PCs. The new approach will require even more powerful signal processing to return a usable 3D image in a reasonable time.


Posted by

Hari Hara Sravan (MGIT ECE 2nd year)