Scientists have found a way to see machines through clouds and fog

It could be incredibly useful to have robots working in a disaster area, autonomous cars circling the city, or having bearing satellites in space, machines that can see through clouds, fog and fog – and scientists could have created an even better system.

The newly developed system works through an algorithm that measures the motion of individual light particles or photons, which are fired from a laser into fast pulses, and are used to reconstruct objects hidden or obscured from the human eye.

What makes the technique even more special is that it can reconstruct the existing light Scattered And jumped around the obstacle along the way.

In the experiments, the laser was able to see objects hidden behind a 1-inch layer of foam.

(Stanford Computational Imaging Lab)

“A lot of imaging techniques make images look a little nicer and a little less noisy, but this is really something we can see invisible.” Says electrical engineer Gordon Weitzstein, From Stanford University.

“It really pushes the boundaries of what is possible by any sensory system. It’s like an inhumane view.”

When the laser light passes through the barrier – the foam, in this study – only a few photons strike the object behind it, and even fewer regenerate it. However, the algorithm is clever to use that little bit of information to reconstruct the hidden object.

Officially, this is called confocal diffuse tomography, although although this is not the first time such barriers have been sought, it offers many improvements – for example, it can operate without knowing how far the hidden object is.

Like other approaches, this system can operate without relying on ballistic photons – these are photons that can travel through a scattering field to a hidden object and without distorting themselves.

“We were interested in filming through the scattering media without these assumptions and collecting all the scattered photons to reconstruct the image.” Says electrical engineer David Lindell, From Stanford University.

“This makes our computer more useful for large-scale applications where there will be very few ballistic photons.”

Large-scale applications such as guiding a self-driving car in heavy rain, or large-scale applications such as capturing images of the Earth’s surface (or other planets) through clouds – there are a lot of potential applications here. Researchers are interested in experimenting with more scenes and more scattering environments.

Current systems are not particularly good at dealing with light scattering caused by fog and fog.

For example, the lidar is clever at detecting objects invisible to the human eye, but the problem begins when rain or fog interferes with its extensive laser scans. Further down, this system may fix that problem.

Before we get ahead of ourselves, we need to keep in mind that scanning using this method can take anywhere from a minute to an hour, so there is still plenty of optimization to work with.

The most impressive achievement is the re-creation of a hidden object in three dimensions invisible to the human eye.

“We’re happy to push this further with other types of scatter geometry.” Says Lindell.

“Therefore, not only objects hidden behind a thick object, but also objects embedded in dense scattering material, is like seeing an object surrounded by fog.”

Research has been published Nature Communications.