RE: Any suggestions for detecting exactly where a ball hits on a 4' x 6' target?
I'm looking to do a similar problem. Once I have the board (I assume plywood, or painters tarp would work), how would I go about using the two sensors to calculate the x/y position of impact?
Any suggested documentation for me to read up / get started on?
Thanks!
I don't know if the cost is in your range or not, but I'd probably use IMU's. Mount one IMU at each impact location and look at the amount of acceleration plus amount of rotation. IMU's further from the impact point will have less acceleration and higher amounts of rotation. Use a calibrated tool like a ball on a string that's dropped from the same height/distance each time to get baseline measurements and do the math from there.
Hi!
it is mainly theory and I didn't test in real worl, but what I have in mind is to get the distance from the ball to the ultrasonic sensor in 2 dimensions.
Imagine only one dimension (let say axis Z) with the US sensor at the top of the target, pointing to the ground.
If your target has a 30 cm diameter, you can expect a detection when you receive a distance measured by the US sensor ranging from 0 to 30 cm, with 15 cm indicating the center of the target. In my proposal, any lenght more than 30 cm would be the ground itself (so you "refuse" the value and wait for next measurment) or the ball hitting lower than the target.
By using a second sensor perpendicular to the first (X axis, parallel to ground), we can imagine you obtain the X,Y coordinates of the ball hitting the target.
This way, 2 measurments of 15 cm (15 cm on X axis AND on Y axis) will be bull's eye!
any other results both ranging from 0 to 30 will be a hit and at least on value out of the range is a miss.
Hope it can help you!
@adriavia - What is the speed of your object? An object moving at 100kph (a good bowler/pitcher) will travel 300mm in 12ms. Light (IR/LASER/LIDAR) might be what you need.
I'm not clear how I would use the IMU in this case. Lets say a board 6ft x 6ft, how would I use the IMU at each corner to figure out where it hit?
OP originally said vinyl so I'm imagining something with some flexibility to it.
Please don't res old threads, start your own. That is why this kind of confusion happens.
Ahh, I see what you are saying... Is the picture below what you are suggesting?
Sensor 1 & 2 would be calibrated for a max range of 6ft and placed at the top of the board and would need a higher FOV to see the entire board. Then each sensor would output a distance from itself to the impact point? Would this work with a board that is 6ft x 6ft?
From there it should be 'easy' to get x,y.
Are there any higher resolution sensors then HC05 SR4? Would I just need an Arduino Uno and those 2 sensors to create a POC? Anything else needed?
I could use vinyl too...sorry!
Yes, 150km/h max, but most likely closer to 100 km/h. Any suggestions on specific IR/LASER/LIDAR sensors to get x,y across a board/tarp/vinyl 6ft x 6ft?
I said it before when this thread was new--use an array of microphones (3 is the minimum, but 4 might be easier), and calculate the landing position when there's an impact. The sound will reach the microphones at slightly different times, and you should be able to figure out a location from that. Note that this is just like GPS, only you're measuring the differential arrival times of sound pulses, not radio signals! (Actually with GPS, there's one sensor and multiple emitters, but it's the same principle.)
Why did nobody mention piezo's ?
Superb triggers and ideal for this type of application.
Also very cheap to implement.
IMU's at each center of impact or zone.
Depending on the accuracy required and material used that could definitely work.
They can be zoned quite easily too.
The program can detect the nearest hit and either ignore the lower readings from the others or, used as an intermediate locator if two hits are close together in the output readings.
Any good suggestions on mic/sensors to use? & some information to read up on?
I ordered a 5 pack of DAOKI High Sensitivity Sound Microphone Sensor Detection Module for Arduino AVR PIC
thanks! will follow along this one and try it to see how I could use it
See post #10
Only did a quick read.
Was surprised it wasn't explored more as it is a pretty common usage for such thing.
Hi All,
My son plays hockey and is also interested in Arduino / technology. My wife was active on another post, but asked I take over 
- link here
We looked at computer vision, but ~100 FPS on a RPi just wasn't enough with the stereo cameras we were looking at, and needed a much more expensive solution.
A hockey net is 4ftx6ft and we're looking to see where his shot went in with x,y coordinates and ideally the speed; and then ideally write it to a DB or something so we can track progress over time & make a bit of a game. Note, the puck is about 1in x 3 inch, and can hit speeds of up to 90mph+.
Based on the other thread linked above, we got a 5 pack of ultrasonic HC-SR04 and a 5 pack of sound detection modules.
We're looking to learn with trial and error here. I wanted to share what we are thinking, and where we are at today. Currently, we have a Uno R3 + starter kit + multiple bread boards in addition to the sensors above.
Basically, we're thinking of trying a tarp covering the net, or some sort of wood, so the puck will make a loud sound on impact. From there / the other post, we want to attempt to use the sound detectors placed at the four corners to calculate the exact ms the sound is detected in each mic and then attempt to triangulate. Today, we got one mic working on sound detection, and would love some additional feedback on how to code / the math for this problem.
The second option we are looking at is the ultrasonic sensors. We also got one working with NewPing, and can see the distance with once sensor. Depending on the FOV, can the sensors see the entire area, we would also place one at each corner to attempt to triangulate based on distance from the sensors. I'm not fully convinced they would have the right FOV / refresh rates to handle 90mph.
Break beams could be an ideal solution, however, I think we would need way to many to cover the entire net given the size of the puck. Are there any 'wide fov' break beams? Otherwise we would need 48 on the Y axis and 24 on the X based on the size of puck.
I know there are other options such as placing multiple piezo into a woven fabric, but it's unlikely we could get the accuracy we would want.
For speed, we were thinking of 2 break beams close to where the puck is shot to calculate the time in between. But the speed is less of a priority at this time as we have a radar gun.
Any other options I'm missing? Does anyone have some guidance on how to use the microphone or ultrasonic sensors to start the triangulation process?
Thanks all!
