A single laser shot through a keyhole may reveal everything in a room

by -
Screenshot: YouTube – Stanford Computational Imaging Lab

techno.rentetan.com – It might be time to cover up your front door peephole for you to worry about privacy.

The skill was once reserved for great heroes to see within a confined room. But researchers at the Stanford Computational Imaging Lab have expanded on a technique termed nonline-of-view imaging in order to see what physical things may be within a single laser light point.

No novel idea is, of course, the non-line-of-sight (NLOS, for short) imaging. It is the smart technology which has been perfected throughout the years in research laboratories to make cameras that can be seen remarkably around the corners and generate photos of objects not visible otherwise or blocked by a number of barriers. The approach has formerly leveraged flat surfaces, such as floors or walls, both in the camera’s and the obscured object’s view. A sequence of light pulses from the camera, usually lasers, rebound off the surfaces and then rebuild the hidden object before finally returning to the sensors of the camera.

Algorithms utilize the information on how long it took to return to produce an image of what the camera could not see. The results are not very clear, but usually enough detailed to establish quickly what the object is.

This is an amazingly smart idea. It might one day be beneficial for gadgets like self-sufficient cars that might identify any risks hidden at the corners long before they can be seen in a car by passengers, thereby enhancing security and preventing obstacles. However, the present NLOS approaches have a major limit: they depend on a huge reflective surface in order to quantify light reflections from a hidden item. It is quite hard, or at least till now, to try to picture what is inside a locked door from the outside.

The imagery technique developed by researchers at the computer imaging lab at Stanford University is so named because it is all needed to see what is inside a closed room is a small hole (such as a keyhole or a peephole), which is sufficiently large to shine through the beam and create a single point of the light inside the wall. As with past studies, laser light bounces off a wall and an object in the room and again from the wall, which finally reflects numerous photons across the hole and the camera which uses a single-photon avalanche photodetector to monitor the time when they return.

Screenshot: YouTube – Stanford Computational Imaging Lab

The new images of keyhole can not calculate just what an object hidden in the room is static. But scientists found that a moving object coupled with pulses of light in a laser gather enough data to form a picture of what it sees during an extended period of exposure. The results are much better than with the previous NLOS algorithms, yet they are still sufficiently detailed to make an educated guess about the size and shape of the hidden object.

A wooden model finally looks like a ghostly angel, but when coupled with a well trained AI picture recognition, it seems highly conceivable to determine that a human object (or object of human design) was in the room.

The research may provide a means for police and military personnel to assess the threats before the door is broken down and stormed in, using nothing more than a small wall breach or a break between a window and a door. It could also offer new ways for the detection of concealed dangers by independent navigation systems long before they are a threat in scenarios where prior NLOS techniques were not practicable in the context of environment.