How Can I Improve My Laser Detection Software for Camera-Equipped Computers?

[X_Art_X]

Hi Guys,
I've felt the need to post a few times in 3 yrs, and this is one of them :)
I tried to go back and thank contributors of another thread, but it was locked
(likely because it was resolved, or 3 years old!).

I'm writing for a camera equipped computer, tracking a red laser beam reflected off a blank wall,
and looking to improve a part of it's software that determines what a laser dot looks like
at current lighting levels, compared to the rest of a wall before it can track it.
This is because it does not just look for the brightest camera pixel, which would be easy,
but the centre of the circular object it sees on the wall, of a colour within a tolerance of
the one gained at startup.

The three colour component values of the colour with the highest red value for a particular frame were actually:
R:0.619608 G:0.576471 B:0.533333

Where 1 is the highest possible value for each component.
I find that red is never much higher in the centre of the beam reflected off a white wall,
but red IS always higher than green, and both of them are always higher than blue.

Is this because red is further away from blue than green in a rainbow?
I'm serious.
Thanks :)

 

[A.T.]

I find that red is never much higher in the centre of the beam reflected off a white wall,
but red IS always higher than green, and both of them are always higher than blue.

Is this because red is further away from blue than green in a rainbow?

It's simply because your laser has the color it has. If you had a purple laser, red and blue would be higher than green.

 

[X_Art_X]

It's simply because your laser has the color it has. If you had a purple laser, red and blue would be higher than green.

Why is the green value always higher than the blue value?

 

[A.T.]

Why is the green value always higher than the blue value?

Not always. It depends on the wavelength of the laser. See picture here:
http://www.philiplaven.com/p19.html

 

[sophiecentaur]

Why is the green value always higher than the blue value?

All three sensors have wide bandwidth (sensitivity curve) and they overlap - they are not really simply red green and blue sensors. The Short wavelength sensor is maximally sensitive in the blue region of the spectrum and overlaps the middle wavelength sensor curve but is pretty insensitive to red. Likewise, the long wavelength sensor curve is highest at red, overlaps the central (green) wavelengths but does not see blue. Green peaks in the middle of the optical spectrum but spills over on both sides. It's these wide curves that allow a three sensor camera to mimic human colour vision. This link is a start.
The red laser was producing an output on the 'green' sensor but very little on the blue because of the fact that the 'green' sensor 'sees it' (by design) and the 'blue' one doesn't..

A;so, did you do your measurements in the dark? Under most lit conditions (particularly tungsten light) there's more of the middle and long wavelengths around and not much short. I reckon the ambient light was polluting your measurements and could have been increasing the proportion of green to blue output.

 

[X_Art_X]

It's easy to test right now, and in the dark at night G & B values are much closer together.
Thanks.

 

[X_Art_X]

Well I never followed up,
That worked by the way, thanks.
Free novel augmented reality game controlled with laser pointer :)

https://itunes.apple.com/au/app/laser-invaders/id595175574?mt=8