My Project

I have a small problem with the idea I have of my Final year project......
I am trying to make something like a *population counter*, it would make use of Infrared LEDs, Two LEDs would be placed per row, and would be made of two columns like in the image shown.... When column one LED is blocked off before column two, it would be registered as an IN, and then it would be added to a counter, but when column two is blocked off before column one it would be registered as an OUT and would be subtracted from the counter......
The challenge I am having now is in the case when a mother and her four month old child are trying to get IN, definitely it would be registered as just one person.....
Please I would appreciate any advice on how to make this project count properly....... I hear MATLAB could do something like that with cameras, I don't know... But i would appreciate any advice on how to go about my final year projects...
All advice would be highly appreciated......

Hi there George,
Your basic idea is good and sound, but the problem of people carrying children, or walking side by side, or too close to each other, is something that would cause troubles even to experienced engineers ! As you suggest, a CCTV camera with computer analysis of images could work, but this would require a complex system with complicated software.
(Although possibly there may be off-the-shelf programs that have already been developed.)

A possible solution for your course project may be to design your equipment without allowing for these circumstances, but mention these possible problems in your submitted notes that go with the project, and give general ideas of how they would be solved.

Depending on the level of complexity you are expected to design to, this may be sufficient to obtain a good project mark, as the examiner should also be impressed by your foresight and anticipation of such possible problems.

I really appreciate your kind words and advice..........

I have considered not allowing the circumstance so many times before, I just thought there was a way ( a simple way) around, I think now with your advice Frank, I would just go on with the project without those checks......

Thanks for your advice, now I am much more confident

Might it be possible to measure the strength of the IR image; & so maybe at least, measure the quantity of suspect readings. Not a solution, but if data were so obtained it would be better than nothing, & maybe could alarm if a higher than expected "suspect readings" were observed.
If your counter could be used in conjunction with another device, for example a turnstile would this not help separate moving bodies.

As this is a project for an electronics course, the complexity of the solution must be geared down to an appropriate level to allow satisfactory project completion in the time available.

(Surely if you have a mechanical turnstile that counts people, you don't need the electronic system at all ?

Perhaps diagonal beams along with the orthogonal beams could be used to distinguish between single humans and groups of humans or humans pushing baby buggies, etc.

If you have, or can obtain some statistics on how often the mother/infant situation happens, you could add a fudge factor to your final pop count.

Frank Bushnell's caveat about side by side walkers is another potential gotcha, but Steve Lawson's diagonal beams would pretty well fix that. More complexity to add to your project design!

I am sorry but I dont even understand what your guys are saying, LOL, But i love the ideas anyway, and I appreciate all, I am going to make research on your opinions, THANKS

Tell us what it is that you don't understand and perhaps we can elaborate.

@steve, your idea of diagonal and orthogonal beams, i dont really understand that, if you think it is a possible solution to my problem, I would like to know it

The beams on your apparatus are orthogonal (to the flow of foot traffic). I'm suggesting adding beams that shine at some angle other than 90 degrees to the flow of traffic. Something like 45 degrees. Thus they would give more information regarding the size or shape of the passing person/object.

But, the more I think about it (and visualize it from different angles), the more I see how difficult it is to allow for all the possibilities: two people side-by-side, perhaps holding hands, or walking in a "side-embrace", fat people, skinny people, people pushing strollers, people carrying babies, people with backpacks, a crowd of teenagers, a mob of kids, ...

perhaps include a row of beams pointing down along the same plane as the beams in your diagram.

And, probably what is needed is a test set up with various beam combinations that is exposed to actual foot traffic so the beam cutting patterns can be captured and studied. And, perhaps some kind of learning algorithm could be used (like the one used by "Watson" -- the IBM computer that was featured on Jeopardy). The algorithm would be fed the beam breaking patterns, along with corresponding foot traffic data.