logoalt Hacker News

IAmBroomyesterday at 1:33 PM6 repliesview on HN

OK, this part was brilliant:

"To avoid this problem, the team divided their 100-milliwatt laser into eight beams. Each beam travels along a slightly different path through the turbulent atmosphere and thus receives a different random phase perturbation. Counterintuitively, this incoherent illumination makes the interference effects observable.

When I first started studying optical engineering, my teacher had worked on the first under-the-RADAR guidance system for bombers. He told lots of amusing stories, like how the pilots insisted on a manual override - so they "agreed" to provide a switch, noting to us manual piloting at near-treetop level and 1,000 ft/s is insane.

He taught us about the nominal amount of turbulence in the atmosphere, and that it limited space-based cameras to about half a foot resolution - a limit he said couldn't be broken. Therefore, license plates would never be readable from space...

Before I was out of grad school, they had broken it with laser techniques on nearby targets. Flash the laser at the same time as the image, scan the laser-illuminated spot, calculate the perturbance, and reverse-filter the image. A lot of processing (for that day), but it could be done back on Earth.

As you can see from the test images, the 8 lasers aren't enough to perfectly smooth out the noise. The noise is probably square-root-8 improved, so resolution should improve by a factor of not quite 3. Move those lasers slightly and repeat 12 times; you've improved resolution by 10. This is easy to do quickly; you should be able to read fine print held by a car passenger on the highway.


Replies

dekhnyesterday at 3:44 PM

We are in the middle of a renaissance of image processing across a wide range of fields. Many of the previous limits are being smashed by using new materials and algorithms. See https://en.wikipedia.org/wiki/Fourier_ptychography for an example

show 2 replies
kevmo314yesterday at 2:57 PM

That's how night mode works on Pixel phones, right? I believe it takes a few images in rapid succession and took advantage of the noise being random which meant a high quality image under a noisy sensor with some signal processing.

show 4 replies
perihelionsyesterday at 2:50 PM

- "Flash the laser at the same time as the image, scan the laser-illuminated spot, calculate the perturbance, and reverse-filter the image"

That's also how some adaptive optics work in astronomy,

https://en.wikipedia.org/wiki/Laser_guide_star

show 1 reply
hammockyesterday at 3:23 PM

>He told lots of amusing stories, like how the pilots insisted on a manual override - so they "agreed" to provide a switch, noting to us manual piloting at near-treetop level and 1,000 ft/s is insane.

You ought to read Tom Wolfe’s “the right stuff” asap if you haven’t already

show 1 reply
quantadevyesterday at 6:48 PM

So what's the summary of how this works? I don't think it was explained well, and I'm fairly up to speed with the physics of photons etc. Is it that the multiple lasers are able to destructively interfere with each other so that they cancel out the noise from each other since the noise will be the same in all of them? That's tricky because if the photons are phase shifted to cancel out the noise that seems like the ENTIRE laser signal would be cancelled out too. Maybe this is what's happening, and the only thing "left over" is the signal from the source (what's being measured)?