djam
Posts: 2
Joined: Wed May 01, 2019 5:09 pm

Sensing an object on a plastic surface

Wed May 01, 2019 5:20 pm

I am a beginner, this is my first post, sorry if it doesn't belong here (please advise where to go for more help).


Here is the video of the thing I am interested to build: https://www.youtube.com/watch?time_cont ... TcIxX_JkfU

I've also attached the screenshot from the video:
Screen Shot 2019-04-15 at 2.25.01 PM.png
Screen Shot 2019-04-15 at 2.25.01 PM.png (204.89 KiB) Viewed 357 times

Project: imagine a smooth plastic surface with a bunch of led lights underneath. A random LED light turns on and stays on until an "object" covers it. Once the light is "covered", it turns off and another one turns on at another random location.
I never worked with sensors and struggling to figure out which one should I use. I believe the one on the video is using magnetic sensor, which makes it a hard requirement for the "covering object" to also be a magnet. Is it possible to NOT use magnets at all, but accomplish the same result? Is there anything similar that I could take a look as a starting point (don't even know what terms to look for)?

Thanks!

blimpyway
Posts: 185
Joined: Mon Mar 19, 2018 1:18 pm

Re: Sensing an object on a plastic surface

Thu May 02, 2019 10:16 am

You could try camera / OpenCV to visually track the puck, specially if it has a certain color not "available" elsewhere in the frame.

And will probably be cheaper than a sensor array solution.

With a laser pointer + two servos you might even be able to dispense with the led matrix and use any slippery flat surface.

-----------------
But if the actual reason of not using magnets is your general lack of experience with sensors, then a computer vision solution could be as challenging or more.

PS the puck does not have to be a magnet, it can be any material and have a small magnet embedded in.

User avatar
omegaman477
Posts: 148
Joined: Tue Feb 28, 2017 1:13 pm
Location: Sydney, Australia

Re: Sensing an object on a plastic surface

Thu May 02, 2019 10:58 am

djam wrote:
Wed May 01, 2019 5:20 pm
I am a beginner, this is my first post, sorry if it doesn't belong here (please advise where to go for more help).


Here is the video of the thing I am interested to build: https://www.youtube.com/watch?time_cont ... TcIxX_JkfU

I've also attached the screenshot from the video:
Screen Shot 2019-04-15 at 2.25.01 PM.png


Project: imagine a smooth plastic surface with a bunch of led lights underneath. A random LED light turns on and stays on until an "object" covers it. Once the light is "covered", it turns off and another one turns on at another random location.
I never worked with sensors and struggling to figure out which one should I use. I believe the one on the video is using magnetic sensor, which makes it a hard requirement for the "covering object" to also be a magnet. Is it possible to NOT use magnets at all, but accomplish the same result? Is there anything similar that I could take a look as a starting point (don't even know what terms to look for)?

Thanks!
I think a simple reflective IR diode/photo transistor pair under each 'spot' would also work. I have got them to sense objects reflectively through a 1mm hole. Machine vision is complex and could be obscured by the player, depending on the environment. The puck does not need to be a a solid magnet, it just needs a small say 10mm rare earth magnet set into it and epoxied in. Either IR or magnets would work fine. Magnets have the advantage of not being false triggered by the hockey stick.
..the only thing worse than a stupid question is a question not asked.

djam
Posts: 2
Joined: Wed May 01, 2019 5:09 pm

Re: Sensing an object on a plastic surface

Wed May 08, 2019 3:21 pm

Thanks for your opinions, really appreciate it. The reason I didn't want to get involved with magnets is it will make a hard requirement to use specific puck and not just any puck. I'll give IR diodes a try, computer vision will be very tough (I think it will require significant processing power + camera(s) and it is a very complex field, for me at least). Again, thanks!

Return to “Automation, sensing and robotics”