Slashdot Mirror
Laser Camera Can See Around Corners
Hugh Pickens writes "Researchers at MIT have developed a laser camera that can 'see' around corners and take pictures of a scene not in its direct line of sight. The camera system fires extremely short bursts of light that can reflect off one object, such as the open door of a room, and then off a second object inside the room before reflecting back to the first object and being captured by the camera, after which algorithms can use the information to reconstruct the hidden scene exploiting the fact that it is possible to capture light at extremely short time scales, about one quadrillionth of a second. By continuously gathering light and computing the time and distance that each pixel has traveled, the camera creates a '3D time-image' of the scene it can't directly see. 'It's like having X-ray vision without the X-rays,' says Professor Ramesh Raskar. 'We're going around the problem rather than going through it.'"
The first customer will either be the TSA or some branch of the military.
High-tech companies would invent anything that would sell to any agency vaguely related to counter-terrorism or warfare these days. If they poured a tenth of the resources they spent developing this kind of devices into finding solutions to the world's real problems, we'd all be cancer-free and solar-powered by now...
"A door is what a dog is perpetually on the wrong side of" - Ogden Nash
Do you know *any* new technology that is not related to an older one ? Me neither.
Given that he's using a femtosecond laser and some kind of exotic "streak" camera, I think he has to essentially raster scan the area he's imaging "around the corner". I doubt the ambient lighting or the scene itself will change in a femtosecond, but the raster-scanner better move pretty fast! A fast raster-scanner might be solvable, even. I think the biggest problem he's going to have is dealing with a non-sparse scene. I think this works well for simple geometries, but once he starts dealing with a really complicated scene, there will be too many possibilities to sort out for each given measurement. (Too many variables!)
I wonder how many real scientific "nay-sayers" he really had to deal with -- in his publication he states that this project is an extension of LIDAR work that was done 10+ years ago. I think it's very cool academic research, but I also think that MIT Media Lab *loves* to hype the every-day/year cool engineering project as though it's a complete world changer.