How can we help? 👋

Camera Positioning

How to position your camera for the best results

Maximum operating distance

  • For overall traffic detection (people counting and dwell time), the maximum detection distance is 40 - 55 feet (12 - 17 meters)
  • For face demographics estimation (age and gender), the maximum distance is 23 feet (7 meters)
  • For attention (viewing vs not viewing), the maximum distance is 30 feet (9 meters)

Lighting conditions

If the lighting conditions are poor, such as a very bright background, the analytics quality may suffer. The cameras we ship have with high Dynamic range (> 100 DB), and should handle different lighting conditions, however, we still recommend avoiding facing cameras to a bright exit as much as possible.

 
Notion image
 

How should I mount my camera?

Visitor InSight w/ content (e.g pairing viewers with screens)

 

Position as close to eye level as possible.

It is best to position the camera as close to eye level as possible with any viewers. This can be accomplished by mounting the camera either on top, or bottom, of the screen by utilizing the camera mount or double sided tape/velcro. Do not position cameras higher than 10 feet.

 

Center the camera on the top or bottom of your screen, the best you can.

This will ensure that you get the most accurate attention metrics to understand which content gets the most views.

Tilt the camera towards your viewers.

If the viewers are looking at elevated screens, such as menu boards, it is better to point the cameras slightly down so that customer faces are better visible for accurate analysis.

 

Avoid view obstructions.

If the screens are behind a counter, such as menu boards, try to avoid obstructions. This can include employees, POS systems, product displays, etc.

 

Real life scenario

In the scenario below, we have a camera and screen mounted at 10 feet high (this is the highest point that we recommend). At this screen height, and with the camera tilted downward, the minimum detection distance needed will be 6’3”. This will capture viewer and demographic data for anyone taller than 4’11”.

If the camera was to be positioned closer to eye level, the minimum detection distance will decrease, demographic accuracy will increase, and will detect a wider range of heights.

 
Notion image
 

What does the data mean?

OTS (Opportunity to see): This would be a total traffic count. In other words, anyone that had an opportunity to see your screens is counted here.

Impressions: This is a count of how many people looked in the general direction of the screen.

Viewing: This is a count of how many people looked directly at the screens. As in the example in the screenshot below, the closer that a person is to a screen, the more likely they are to be a viewer.

Notion image
 

Visitor InSight Standalone (dome cameras, cameras not paired with screens)

From (link):

  • If you are mounting the camera separately from a screen (Visitor InSight Standalone), customers may be viewing anywhere. Ceiling-mountable cameras looking straight down, often just see the top of the head, so demographic analysis is usually less accurate. If demographics are important, we recommend mounting cameras to minimize the vertical angle to see as much of the face. This often means either as close to eye level as possible on a wall, or further away and zoomed in like across the room looking at your entryway. As a general rule, we prefer vertical angles under 30º for the best demographic results, although body counts and dwell will continue to be accurate above.
 

See Screenshots 4 and 5 for how the camera can be angled when implementing the AI-Standalone solution.

 

Screenshot 4: minimum distance from camera for face detection (4’11” - 6’5”)

Notion image

Screenshot 5: minimum distance for body detection (4’11” - 6’5”)


Notion image

For the kiosk scenario or self standing screens where the camera is mounted close to head height, here is a range of angles depending on how high the camera is mounted. The distance from camera is 25”, close to an arm’s length:

Notion image
 
 
Did this answer your question?
😞
😐
🤩