Long-range surveillance that can see through turbulence in real time.
“We’ve succeeded in getting super-resolved imagery out of bad turbulence… Adaptive optics involves changing a mirror or something that delays the light in different parts of the beam, if you get that change right you can get that nice clean image back, so you undo the turbulence in real time.”
Turbulence in the atmosphere is a major barrier to long-range surveillance. Dust, heat haze—these have long been challenges for a camera attempting to capture a clear image at long range.
Associate Professor Andrew Lambert, from UNSW Canberra’s School of Engineering and Information Technology, works on not only correcting for the effects of turbulence caused by distortions like heat haze, but also on capturing and recording the effects of the turbulence itself.
“In the early ‘90s we established what most people thought was ridiculous, which was to look through heat haze with cameras,” said Dr Lambert.
He explained that the method works by taking short enough exposures to effectively “freeze” the effects of the atmosphere. Each exposure captures a slightly different distortion, which can then be programmed out of the imagery using digital post-processing.
But without significant computing power, this couldn’t be done quickly enough—that has changed with advancements which have allowed for post-processing in real time.
“So, two things happen if you’re successful: you get a nice image and you get it fast, and you get a visualisation of the distortion—it’s like seeing the wind,” said Dr Lambert.
“You can use this for aerodynamics and Defence interests, for example anti-missile work. That information is really useful. The next step in this is, if you’re able to do this really well, you get super resolution—an image which is sharper than what you would have got on a beautiful clear day.”
The challenge with super resolution is, again, the time it takes to process the imagery. To get super-resolved imagery fast Dr Lambert looks to “adaptive optics”—the process of changing the light as it comes in rather than only processing at the end.
Adaptive optics involved changing something, like a mirror, that can delay the light in different parts of the beam of light coming into the imaging device. This also means being able to read the turbulence at the time the image is being taken and then correcting for that.
“If you get that change right you can get that nice clean image back, so you undo the turbulence in real time,” said Dr Lambert.
“If you do that with some of the fast processing afterwards, you can come up with some really neat instruments.”
Dr Lambert compared the technique to similar processes used in astronomy—however, in this case, the degree of difficulty is dramatically increased due to the severity of the turbulence.
The final challenge is combining these techniques in a device small enough for people to use, or to put on an unmanned aerial vehicle (UAV).
“That research has led to my participation in a NATO Science, Engineering and Technology committee for mitigation of the effects of turbulence on imaging in long-range communications.”