Improving the Laser Dot Rendering in The Refraction Simulator

Mick West

Administrator
Staff member
Something I'd like to improve is the rendering of the laser when it hits the camera. Currently what is being rendered is the diffuse beam (i.e. the beam light that is scattered by the air.) But when that beam covers the camera, there should be a very bright spot of light at the position of the laser source.

I did a few experiments in daylight, shining a laser at a tree about 330 feet (100m) away.
Metabunk 2019-02-24 13-52-16.jpg
From this distance, it was just visible to the naked eye. Up close there was a visible spot 2-3" across
Metabunk 2019-02-24 13-54-08.jpg

Iphone X camera looking back at the laser shows a clear bright light that grows and shrinks as you move in and out of the dot.
IMG_4946-stab.gif

Metabunk 2019-02-24 14-14-32.jpg

So I kind of need something like this in the simulator.
 
Last edited:
THinking about the spread of the beam raises some interesting things. Here I added three laser beams to the simulation, one angled up slightly at -0.056° (the green one) and then two that are just 0.001° on either side of this.
https://www.metabunk.org/refraction...0Front~multiple~0~gap~0))~localDefault~true)_
Metabunk 2019-02-24 15-20-47.jpg

If we remove refraction:
Metabunk 2019-02-24 15-27-46.jpg

So the refraction is causing the beam to diverge a LOT more than you would expect, which in turn makes it a lot more visible from a wider variety of positions.

The spread I observed was about 2" in 330 feet, so about atan(2/(330*12)) = 0.029° spread. Say 0.03° If we Stick that in for the upper and lower blue beams, we get:

The top edge now being 60 feet above the center with refraction
Metabunk 2019-02-24 15-34-55.jpg


Without refraction, the divergence is far less, about the 20 feet I estimated earlier.
Metabunk 2019-02-24 15-36-05.jpg
 
Early morning experiments for better light:

Right next to the laser, the beam is about 1mm, but there's plenty of scattering quite widely.

Metabunk 2019-02-25 06-47-28.jpg


330 feet (100m) away

Laser's perspective
Metabunk 2019-02-25 06-52-16.jpg

Target's perspective
Metabunk 2019-02-25 06-50-50.jpg
Size of dot (card is 12.5 inches, 32cm across)
Metabunk 2019-02-25 06-49-38.jpg

The colors are:
  • Solid white
  • mixed solid white and solid green
  • solid green
  • scattered green
Which you don't see with the naked eye, but are really artifacts of the camera.

Looking back at the laser, with the camera exposing video for the dim light, we get more intense "dots"

Metabunk 2019-02-25 07-00-19.jpg
Metabunk 2019-02-25 07-00-38.jpg
Metabunk 2019-02-25 07-01-03.jpg
Metabunk 2019-02-25 07-01-33.jpg

Metabunk 2019-02-25 07-02-03.jpg
 
Using some laser safety glasses allowed me to see the size of the business portion of the beam more easily.

Metabunk 2019-02-25 07-04-13.jpg

At about 40 feet:

[compare]
Metabunk 2019-02-25 07-05-36.jpg
Metabunk 2019-02-25 07-06-13.jpg
[/compare]
 
Of nerdy interest, on the target the solid white center has a ring of cyan (green+blue) between it and the solid green
Metabunk 2019-02-25 07-10-13.jpg

But a few seconds later in the same shot, facing the laser, it's a ring of yellow. (green+red)

Metabunk 2019-02-25 07-11-13.jpg

This must be something sensor specific, to do with how the red, green, and blue sensors in the camera get overloaded. In reality, it's all green.
 
So, I think for the improved rendering, I need to trace a center ray, and a top and bottom ray, and then see where they hit the plane of the camera, then interpolate (and possibly do a binary search to refine) to get the incident angle of the direct path, and hence the position of the center of the dot
Metabunk 2019-02-25 07-20-11.jpg
In the above, the camera is on the left. The laser is on the right, the green line is the nominal center of the laser, and the red and blue indicate the spread.

The orange line (with the mouse cursor) is the line of sight from the camera to the laser, this will need a binary search to narrow down. This involves narrowing the extents of the beam depending on which half it is in. So here I'd cast a new laser ray midway between green and blue, and then see which side of that ray the camera is on. This process will also give me the angle from the center beam, and hence a more accurate basis for brightness.
 
The spread I observed was about 2" in 330 feet, so about atan(2/(330*12)) = 0.029° spread. Say 0.03°

0.03° is 1.8 arc minutes, or 0.5 mRad (milli-radians)

Other sources:
https://www.quora.com/What-is-realistically-the-least-possible-beam-divergence-of-a-laser
For reference a typical laser pointer has a divergence angle of 1 to 2 milliradians (about 5 arc minutes).
Content from External Source
mRad seems to be the most common unit for giving beam divergence. Here's a bunch of
https://www.wickedlasers.com/laser-tech/laser_beam_comparison.html
Metabunk 2019-02-25 08-14-10.jpg

Similar here, mostly 0.5 to 1.5, so I think a default of 1.0 is good. I'll make it editable.
http://www.lasersafetyfacts.com/resources/FAA---visible-laser-hazard-calcs-for-LSF-v02.png
Metabunk 2019-02-25 08-52-17.jpg

Since I was measuring in daylight, I got too small a value. Looking at the dawn shots, it was probably three times as much, or 1.5 mRad (0.08°).

My laser is a generic 301 green laser (532nm). The beam can be focussed, and I had it on the tightest beam. This review says that's about 1mRad, consistent with my observations.

https://laserpointerforums.com/threads/dx-sdlaser-301-50mw-focusable-new-pics-divergence.51631/
Divergence
When focused to infinity ( minimum ) the divergence is 1mRad, what I think is a good value.
When focused to burn ( big dot) the divergence is 22,7mRad.
Content from External Source
 
Last edited:
Back
Top