we have a project which involves detecting a (heavy) ball being dropped by a person,
and are thinking of the kind of sensor we could use to accurately detect it.
Note that the ball is being dropped in a known 3D zone,
so we know "where" to track motion.
We'll be outside so we must handle weather changes (lighting conditions, etc...)
We want to detect the ball before it hits the ground,
so we cannot rely on a sensor reacting to the ball's impact.
We also cannot use load cells to detect weight loss as the ball is being dropped
as it would interfere with other components of the project
(we have a fixed structure and we do not want to modify it to insert load cells)
So we are left (as far as i know) with the following options:
PIR (likely not fast enough and unable to accurately discriminate between the "volume of interest" and the outside world)
Sonar (ikely not fast enough and unable to accurately discriminate between the "volume of interest" and the outside world)
IR Distance (this one may work, we'd have to check the sensor update rate though)
Camera (using a computer vision toolkit, pretty confident it would work but maybe overkill)
Laser tripwire web (pretty confident it would work but lots of hardware for a dense enough web)
Is there any options we could include here?
Which one makes most sense in your opinion?
Thank you!
PIR will work if your ball is hot (i.e. emits IR light). But normally the ball will be at ambient temperature, so invisible to a PIR.
How big is the ball compared to the detection zone? How far from the detection zone are your sensors? Is there any demarcation of the detection zone, e.g. a ring the ball has to go through? Maybe magnetic sensing is an option. Or if conductive, capacitive sensing could work.
If the ball is big compared to the zone it passes through, you may need just a few light beams for it to be detected. Modulated IR is probably the most reliable here, but rain can of course mess it all up. A few beams in parallel, a little closer than the size of the ball, and you're guaranteed to break at least one of the beams.
The ball is like a bowling ball,
thrown in a volume like a cylinder,
diameter 2 meters, height 3 to 5 meters
Our sensors will likely end up on the inner surface of the cylinder,
at mid height (before the ball hits the floor but after its velocity direction stabilizes)
So we could have some sort of ring inside the cylinder, yes!
What about the ball do you need to detect? That it was dropped at all? Predict where it's going to land?
IR beam-break sensors will be best. What they show in the movies with lasers but real alarm systems don't actually use lasers.
With a bit of skill, a small number of emitters and receivers can make many 'beams'. If receiver 3 is seeing the unique digital pattern emitted by transmitter 1 then you know that 'beam 1-3' is unbroken. Think like many TV remotes each with a different button pressed.
From the point of view of the Arduino, the falling ball is very slow. It can easily send several detection pulses in the time that it takes for the ball to fall a centimeter.
wvmarle:
PIR will work if your ball is hot (i.e. emits IR light).
You may confuse near IR (which is light) and far IR which is heat radiation and is not normally called light.
Best to make a clear distinction between them. Shiny metal objects do not emit heat radiation, merely
reflect it from nearby objects/surroundings BTW.
Warm/hot objects will emit heat radiation and be easily visible in a suitable thermal camera, but that will not help
you detect the object before it hits the ground probably (non-military thermal cameras are frame-rate
limited to 9fps max).
CV might be the best way if the balls are distinctively coloured.