I agree with Ben.<br><br>A grid of sensors (in the ceiling and along the sides), perhaps including some motion sensors should be enough.<br>I think that this could be arduino controlled, requiring some external shift registers depending on how many sensors you use.<br>
Conceptually it should be too difficult for you to build a model with only a few sensors in order to prove the proof of concept.<br><br>
Regarding longevity you want to avoid if possible using something like
openkineck or opencv, basically anything that requires a PC or laptop
doing the work. That's bound to be stolen or broken.<br><br>Cheers<br><br>Martin<br>
<br><div class="gmail_quote">On 26 November 2010 22:33, Ben Dembroski <span dir="ltr"><<a href="mailto:ben@dembroski.net">ben@dembroski.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Fun project!<br>
<br>
Depending on the budget, I would think that an overhead array of sensors<br>
could be used to track people's movements using simple shadow detection.<br>
You wouldn't need to use colour tracking. By tracking the movement of<br>
the people against a grid of sensors. If the sensor's are triggered<br>
sequentially in any direction, that's likely a person moving. Off the<br>
top of my head, I'd think you'd need to track movement in at least 2<br>
axes to prevent the system getting confused.<br>
<br>
It might be possible to do this without the complicated hardware of<br>
something like a PC doing the image detection -- which would be the weak<br>
point, longevity wise.<br>
<br>
Just two off-the-top-of-my-head pence worth ;)<br>
<font color="#888888"><br>
--<br>
Ben D.<br>
</font><div class="im"><br>
<br>
On Fri, 2010-11-26 at 21:52 +0000, Clive Mitchell wrote:<br>
> That's going well into the realm of image recognition, which would be<br>
> even harder as the person would change colour when their "wave" of<br>
> light commenced.<br>
><br>
> I think a wave of colour triggered by a pedestrian and moving at a<br>
> constant average walking speed might be a much easier option. This<br>
> could also be implemented in a shift register format. I presume that<br>
> the system will detect people entering at either end?<br>
><br>
> On 26 November 2010 14:54, Roy Mohan Shearer <<a href="mailto:fy81le@googlemail.com">fy81le@googlemail.com</a>> wrote:<br>
> > The<br>
> > effect that we're going for is that as each individual enters the<br>
> > underpass, the spotlights increase in intensity and a colour is<br>
> > assigned to that person which accompanies them as they walk through.<br>
><br>
><br>
><br>
<br>
<br>
<br>
</div><div><div></div><div class="h5">_______________________________________________<br>
members mailing list<br>
<a href="mailto:members@electronclub.org">members@electronclub.org</a><br>
<a href="http://lists.electronclub.org/cgi-bin/mailman/listinfo/members" target="_blank">http://lists.electronclub.org/cgi-bin/mailman/listinfo/members</a><br>
<br>
Instructions for changing your mailing list settings:<br>
<a href="http://lists.electronclub.org/emailhowto.html" target="_blank">http://lists.electronclub.org/emailhowto.html</a><br>
</div></div></blockquote></div><br>