A new vision system reconstructs particle movements with the help of an algorithm, being able to see through clouds, fog, and other structures.

 

Scientists Highlight which to the human eye is like seeing through walls.

Researchers at Stanford University have developed a Innovative vision system that reconstructs scenes hidden three-dimensional thanks to the movement of photons or elementary particles of light.

 

It could be great Usefulness for driving autonomous vehicles in emergency situations. fog or heavy rain, or even to improve the visualization of Satellite imagery in foggy environments.

In a statement, the specialists they stressed that this mechanism runs the frontier of what can be can be achieved with any vision system.

Seeing objects through clouds and fog | Stanford News

The system employs a software specially created for this purpose, which works integrated into an algorithm to be able to reconstruct Three-dimensional objects hidden in images.

 

To do so, it requires some location data and a few photons as information.

With the goal of seeing in environments that scatter light in multiple directions, the system pairs a laser with a photon detector of extreme sensitivity.

 

The laser scans the structure that obstructs vision, such as a foam wall, but in at some point, one of the photons in action will pass through the foam and it will bounce off the hidden objects you want to display.

Subsequently, the photon will pass through the foam again, but this time in the opposite direction, and it will reach the detector that owns the system.

 

That's when the software and the algorithm kick in, decoding the information provided by the photon or elementary particle of light.

 

When you finish your work, they will offer a reconstruction of the hidden objects that can be visualize in three dimensions.

 

The highlighted that this technique complements other vision systems that achieve similar results when seeing through structures and visual barriers that seem impregnable, but that they actually do it on a microscopic scale, usually for applications in medicine.

The new system, on the other hand, is geared towards large-scale situations. such as driving autonomous vehicles in situations heavy fog or heavy rain.

 

It is also believed that it can be operational to optimize the display of images satellites from the surface of the Earth or other planets, which have a hazy atmosphere that makes it difficult to see.

The infrastructure required to develop the new system is not too complex.

 

Only some hardware components a bit more advanced than those that can currently be found in self-driving cars, with in order to achieve a good response from the software and the algorithm.

Depending on the luminosity of the objects hidden from the display, the scan takes between a minute and an hour, so except in the testing phase of the device.

 

An extra advantage is that the algorithm reconstructs the hidden scene in real time, and also the application can be run on a medium-sized laptop Benefits.

 

Three-dimensional imaging through scattering media based on confocal diffuse tomography - Lindell, D.B., Wetzstein, G. Nature Communications (2020).