Testing People Traffic Counts

SOLVED
Ericka
Here to help

Testing People Traffic Counts

Is anyone using the MV cameras to count traffic entering a building?

If so, what did you do (if anything) to feel confident that the numbers produced by the cameras were reasonably correct?

1 ACCEPTED SOLUTION
NolanHerring
Kind of a big deal

Have not done this, however I suppose if the amount of traffic is reasonable, one could take the time (maybe increase video playback speed to 4x) and count them manually, and compare the results with the automated system.
Nolan Herring | nolanwifi.com
TwitterLinkedIn

View solution in original post

14 REPLIES 14
NolanHerring
Kind of a big deal

Have not done this, however I suppose if the amount of traffic is reasonable, one could take the time (maybe increase video playback speed to 4x) and count them manually, and compare the results with the automated system.
Nolan Herring | nolanwifi.com
TwitterLinkedIn

View solution in original post

That is a great idea, thank you Nolan!
PhilipDAth
Kind of a big deal

I have a customer that does this.  They haven't reported any issues.

We've found the people counting with MV to be inaccurate on the high side. The feature is still very useful to trend busy hours/days, but I wouldn't consider it accurate. In busy environments with people walking behind each other or behind obstructions, the inaccuracy increases.

 

Meraki does a good job of explaining the things that can throw off the count here:

 

https://documentation.meraki.com/MV/Video_Analytics/MV_People_Detection

 

We've seen the accuracy improve over subsequent releases and I fully expect it to continue to improve.

Don, 

thank you for taking the time to reply.  How did you know that the count was on the high side?  What were you comparing it to?

We have dedicated people counters in some of our retail stores located above the entrances and the MV counts are much higher than what the people counters are reporting. 

 

http://www.brickstream.com/NewDatasheets/FLIR-Brickstream3D.pdf

 

Don, thank you again for taking the time to reply, and to send the Brickstream link.  We are also considering more traditional traffic counters, so I'll look at this one.

 

any idea of the difference, percentage-wise, between Brickstream and Meraki?  

 

Do you think it's because the Meraki camera does not account for directionality?  Are you getting the traffic count from the dashboard, or pulling it via API?

 

We're just comparing to the dashboard numbers. I have not done a recent comparison. When this was first launched (maybe a year ago?, older firmware) we were seeing Meraki numbers 10 times higher than the traditional people counter.

 

We have seen improvement over time and our store we used for comparison was extremely busy (lots of people dwelling and moving around each other and other obstructions). This is a very difficult use case for the MV camera.

 

We still love the people count analytics as a comparative measure. In 3 clicks i can get to the busiest minute, in the busiest hour, on the busiest day of the month and immediately see what the store looks like at max capacity. We find that extremely valuable.

 

If dead accurate people counts are important to you then there are more traditional options that may be better suited.

Phillip, did you mean your customer uses MVs to count traffic, or that they have done a test to verify the traffic counts? Thanks for taking the time to reply!
PhilipDAth
Kind of a big deal

>Phillip, did you mean your customer uses MVs to count traffic, or that they have done a test to verify the traffic counts? 

 

They haven't done any test to verify the counts.  They use them to count people both at the entrance at at various places in the stores.

Phillip, do you know how they account for directionality? or do they just take the count as-is from the APIs?  because that is a count of people entering the camera's frame of view (so it wouldn't differentiate if they're going in to or out of the door).

PhilipDAth
Kind of a big deal

The cameras count everybody walking through the frame.

 

So if your entry is also your exit you'll need to halve the number.

Ahoste
Getting noticed

Hi,

 

I'm currently trying to replace our traditional 'above the door' counters with the MV12 or MV32.

I'm trying to setup the API calls to populate our dataserver with the count. 

 

I just want to receive the door count but I only see the MV sense API calls, so does one need the MV sense license just to be able to get this simple statistic or is there another API call I can make to get this data?

 

Thanks already!
cheers

PhilipDAth
Kind of a big deal

>so does one need the MV sense license just to be able to get this simple statistic

 

Yes.  Note all orgs get 10 of these licences for free (you already have them, just need to assign them to the cameras).

Get notified when there are additional replies to this discussion.
Welcome to the Meraki Community!
To start contributing, simply sign in with your Cisco account. If you don't yet have a Cisco account, you can sign up.