I’m looking to build 2-4, maybe more, people counters. Some means of counting individuals passing through our front doors, sometimes they come in in a steady stream, so it needs to be fairly discerning. It also needs to be able to show a total count/possibly resettable, should be self-contained power, ie 9v battery, although if we install overhead I may be able to get a power drop, but then it would be difficult to read numbers without using a ladder, unless the sensor/power unit can be separate from the total count display or connected with a small wire(?), and ultimately it needs to be inexpensive (CHEAP) as I’m footing the cost myself. I tried searching for help on the internet (saw webpages for the Vellman IR Light Barrier with abilities to notify when visitors enter - not count (Radio Shack - Their Tech Support -TED- was no help whatever) then saw the “Redboard - Breadboard” post from Sparkfun and saw the little bit of info on the Arduino Controller. I’m not totally inept, but its been quite a while since I tried building anything, (My previous experience was PRE-IC Chipsets) but if anyone can at least point me in a direction or two, I would greatly appreciate the help
The actual count operation is not difficult; it can be done with any Arduino or other processor. I produce a few types of off the shelf products that handle event counting and report the data to a PC if you are interested.
However, bear in mind that it may not be as cheap as you think to DIY: by the time you’ve added up all the parts to make a people counter similar to a commercial unit, you’re not far behind in cost (if at all). Where DIY really shines is that you can be as flexible as you want.
Where you want to start is deciding exactly what you are detecting: total number of activations of a sensor (e.g., people at a doorway crossing sensor, or a door switch that detects when the door opens) or entries/exits (you count up when people enter and count down when they leave: this requires two sensors and logic).
Next, what do you want to do with the data? Show it on a display is enough, or does it need to be stored on a PC or a database?
Thank you, Yes you have a very valid point regarding the cost to build vs. buying ready made. I noticed from my searching this far that the Vellman IR Light Barrier ws $11 or $13 per unit, and the Redboard Breadboard for the Arduino was $25. I haven’t looked at the cost for the Arduino Controller recently but I thought they were about $100-$150 a year ago or maybe that was Rasberry Pi. I just had a thought, the one Arduino Controller can accept inputs from 2 or 2 sets of sensors from the 2 front doors, right? But let me give you some additional specs and maybe you can give me a price. We have 2 sliding doors that allow both ingress and egress, and when people are coming off a bus they tend to stream in, not quite as much streaming out, therefore I need a sensor that can discern a very brief gap between people (unfortunately we don’t have turnstiles and one-way movement). We’ll just need to run some samples to determine what the error factor is and when we put the numbers into a Db let the computer make some realistic adjustments. The other factor we need to consider, though not you, is we also have a set of double sliding doors leading out onto a rear deck, from which people can and do just walk around the building from the deck, many not re-entering the building to exit through the front doors, and some even entering the building from the rear. Then there are 2 additional manual doors the allow both ingress and egress. So basically I want to be able to be as accurate as possible with the count in and out of the front door, but then we’ll factor in an error coefficient for various situations. I’d like to be able to read the numbers every hour to enter the value into a spreadsheet or database on our computers to track the customer volume. (We currently run numbers as high as 800-900 people per hour generally in the summer months with as many as 16-17 busses and 60-150 people per hour in the winter months (low counts of 8-19 per hour in the early morning and late evening hours) I just would like to dispense with the manual clickers we currently use and BTW, even now we really never get all the people entering/exiting from all the other doors. As in my initial post, if it winds up bankrupting my personal discretionary cash fund, I’ll have to forego the whole thing.
I don’t know if you need to count every visitor accurately or just get a better estimation than you have now. Let’s say you double count because a single person entering and then exiting counts as two people. Then you can just divide the total count by 2 and you’re fine. Likewise, unless a large percentage of people are exiting through a different door, that probably won’t throw off your count that much.
My first suggestion would be to experiment. Get one of those people sensors and a single Arduino (they’re in the $30 ballpark, not $150), build some counting software and see how well that works for a single door. That will get you a lot of data and experience to work with. At the count rates you’re looking at, a single Arduino can handle many input sensors.
As far as cost, I find that people tend to forget about things like power supplies, enclosures to house circuitry, connectors and wiring. That can add up very quickly, so be sure that you budget for that.
You can continue to post your progress here to get more help from those of us in the peanut gallery, or if you want more one on one discussion you can contact me through PM and we can email. I like discussing people’s applications and I understand that they often don’t want to talk about what they’re doing in “public.”
As you've stated the problem is getting an accurate count given the "streaming". Let ask ... do the hordes come in side by side as well as packed, front to back, through a single doorway ? I was wondering if some pressure pads on the floor might be a more accurate way of separating one person from the next ?Bob the Builder:
We have 2 sliding doors that allow both ingress and egress, and when people are coming off a bus they tend to stream in, not quite as much streaming out, therefore I need a sensor that can discern a very brief gap between people (unfortunately we don’t have turnstiles and one-way movement). We’ll just need to run some samples to determine what the error factor is and when we put the numbers into a Db let the computer make some realistic adjustments. The other factor we need to consider, though not you, is we also have a set of double sliding doors leading out onto a rear deck, from which people can and do just walk around the building from the deck, many not re-entering the building to exit through the front doors, and some even entering the building from the rear. Then there are 2 additional manual doors the allow both ingress and egress.