Difficulty: starter
Content
| Table of Contents | ||||
|---|---|---|---|---|
|
Learning Objectives
After reading this article, you’ll be able to:
Understand PointGrabXOVIS's people detectioncounting
Define if PointGrab’s XOVIS’s data types suit your use case
What is
PointgrabXovis?
The PointGrab™ sensors detect people’s presence, locations and numbersXOVIS sensors detect people flow across a(n imaginary) traffic line, and send analytics data over a secure, encrypted network connection.Desk, Meeting Room and Reception areas are intended for real-time people counting. They produce information such as “Desk A is now occupied” or “Meeting Room B now contains 7 people”.
On the other hand, Traffic Line areas are intended to track traffic going in and coming out of a specific location. They produce information such as “In the last update period, 5 people crossed the Traffic Line area”going IN to the area, 3 people crossed the line going OUT”. The delta, or difference, between these two values provides us with the information on how many people are still in the tracked area (utilization).
How does it work?
Non-intrusive solution – images are never transmitted or stored on the sensor
All of the image processing is done on the sensor itself, the sensor does not store or output images, and has no interface to extract images, so privacy is fully protected. The sensor will only transmit the analytics data to the cloud management system.
As the sensor tracks peoples’ movement rather than personal devices (e.g. mobile phones), it is completely anonymous.
In the screenshot, you can see how the coordinates of the people arrive at Pointgrab? These are transmitted to the Workplace IoT platform as "2 people within detection range of sensor”.
The functionality of Xovis 3D stereo vision sensors is based on two optical CMOS image sensors arranged at a certain distance to each other. The images are compiled to a stereo image in real time on the device. The processed images are not stored.
Pixel lock
All Spacewell Xovis devices come with Pixel Lock: an irreversible function that applies a filter to all scene images right after the embedded processing.
People can be clearly seen, but not recognized.
Documents or screens are impossible to read.
Edges are kept sharp, to allow accurate placement of geometries or stitch sensors for wide entrances.
It is not possible to reproduce a sharp image anymore based on the Pixel Lock image.
How often is data captured?
Technically, the sensor can check every 15 sec – 15 minutes for coordinates. It will count the people in the area for at least half of the time of update timeis always on and could update the Workplace platform as soon as a detection has occurred.
As a default, the data is uploaded into the Workplace Cloud every 1 min: this interval is low enough, that a person in the room will be captured, yet it also reduces false detection issues.
Decay Configuration options
Perhaps you would want to indicate in Workplace Live Occupancy View that if a Desk sensor has been triggered, the location is occupied for at least 2 more hours, so that colleagues and guests can go to lunch or drop by a meeting without having to fear that someone is clearing out their workplace? But at the same time you want to indicate that a meeting room has freed up as quickly as possible, so that it turns green (“available”/”not occupied”) and people can recognize this at a glance?
The Workplace back-end Studio configuration options have you covered: per room category, both the icon and the decay timer can be modified.
| Info |
|---|
Knowing that Xovis data is processed every |
minute |
if there is a value change, this change is documented
if there is no value change compared to the last message, the data is stored after max 10 minutes. This means that in Workplace, you might see “last update from the sensor received 4 minutes ago”
Pointgrab Accuracy
Pointgrab provides highly accurate data on Presenceand allowing for some glitches, the timer can be set to anything higher than 5 minutes (5min being the minimum). |
How is data treated?
When a XOVIS device is configured using Footfall, the footfall data will be visualized on https://spacewell.atlassian.net/wiki/spaces/KB/pages/159973452/Live+Data+floorplan#View-by Footfall IN/ Footfall OUT + Footfall TOTAL.
The difference between Traffic IN and Traffic OUT indicates how many people are still inside the location. This value is calculated and is used as (indirect) Utilization value. (see https://spacewell.atlassian.net/wiki/spaces/KB/pages/492545/Utilization+Occupancy+Matrix#Indirect-Utilization-&-Occupancy)
The calculated/ indirect Utilization value
real time data can be visualized in the Live Floorplan View By Space > "Utilization".
historical data can be visualized in the Replay - Timelapse and in the Space monitor dashboard.
In theory, the traffic line counter just continues counting. Because of the use cases within the Workplace platform (know daily flux of people coming in and going out; and indirect utilization), counters are reset at midnight.
XOVIS Accuracy
XOVIS provides highly accurate data on Traffic Line detection, yet some factors might impact people detection:
Make sure lights are turned on/there is enough daylight. It should not be too dark in the room
Nothing should be blocking the sensor detection area
items might be suspended from the ceiling (eg HVAC or light fixtures)
items might be too high and block visibility (eg high back chairs or office screens)
Reflections in windows (might be avoided by restricting count area in the configuration)
Installation needs to be level (not tilted)
Installation location: make sure to verify measurements (height + length/width) in the config platform
Find more information on
https://www.pointgrabxovis.com/our-product/
https://www.pointgrab.com/wp-content/uploads/2018/04/CogniPoint-Data-Sheet.pdf
Search
| Live Search |
|---|