On site, mounting height is often chosen by whatever is easiest to reach: the ceiling, under the eaves, an existing pole. The recording shows the result: the top of a head instead of a face, a zone under the camera where nobody is visible, a camera whose direction can be turned with a stick. Height and tilt angle directly determine all of these. Below we work through image geometry from the perspective of a human viewer with simple trigonometry, and give starting values by area type and a post-installation check method.
How height and angle affect the image
Three geometric variables determine a camera’s image: mounting height (H), horizontal distance to the target (d) and tilt angle. Within this geometry, the lens determines how much width is seen; the calculation is in the camera lens and focal length calculation article.
As height increases:
- People block each other less in crowds, and general situational awareness improves.
- Reaching the camera by hand or with a stick gets harder.
- The real distance to the target grows; pixel density falls for the same area.
- The view of vertical surfaces such as faces and plates gets steeper; detail is lost.
- The unseen zone under the camera grows.
- Maintenance may require a platform or personnel lift instead of a ladder.
Height is a trade-off. If both general monitoring and identification are expected from the same camera, one will always be weak. You can’t have both at once.
Tilt angle and face visibility
The value that determines face visibility is the angle between the horizontal and the line drawn from the camera to the target’s face. Taking average face height as about 1.6 m:
Vertical angle = arctan( (H − 1.6) / d )
H = 2.5 m, d = 4 m → ≈ 12.7°
H = 3.0 m, d = 5 m → ≈ 15.6°
H = 4.0 m, d = 5 m → ≈ 25.6°
H = 6.0 m, d = 5 m → ≈ 41.3°
H = 6.0 m, d = 10 m → ≈ 23.7°
As the angle grows, the vertical size of the face shrinks in the image; the forehead, hair and cap peak take the face’s place. Installation guides for cameras with face capture usually give an upper limit of around 15° for the vertical angle; use the value in your own model’s guide. For a human viewer, a slightly larger angle can be accepted for recognition purposes. Once the angle exceeds about 30°, the face of a person walking with their head tilted forward largely disappears from the frame.
What the examples show: there are two ways to reduce the angle, lowering the camera or moving it away from the target. Moving a camera at 6 m height 10 m away from the target brings the angle down to about 24°; but then the distance exceeds 11 m, and a longer focal length lens is needed for the same pixel density. At identification points, the most efficient solution is usually to place the camera 3–6 m behind the door or turnstile at a height of 2.2–3 m and watch the passage point with a narrow frame. The common mistake is the exact opposite: a camera mounted on the ceiling right above the door that sees only the tops of the heads of those passing through.
How blind spots form
A camera can’t see the area viewed at a steeper angle than the lower edge of its vertical field of view. This leaves a blind zone under and in front of the camera:
Lower edge angle = tilt + VFOV / 2
Upper edge angle = tilt − VFOV / 2
Blind zone (ground) = H / tan(lower edge angle)
Blind zone (head height) = (H − 1.6) / tan(lower edge angle)
Far limit of view on the ground = H / tan(upper edge angle) (if upper edge angle > 0)
Example: H = 4 m, VFOV = 36°, tilt = 25°
Lower edge = 43°, upper edge = 7°
Blind zone on the ground ≈ 4.3 m
Blind zone at head height ≈ 2.6 m
Far limit on the ground ≈ 32.6 m
In this example, an object on the ground within the 4.3 m right in front of the camera isn’t seen; the head of a person coming closer than 2.6 m to the camera leaves the frame. Reducing tilt extends the far limit but enlarges the blind zone. Increasing tilt shrinks the blind zone but cuts off the far area. If the upper edge angle drops below zero (the camera looks at the horizon or upward), the top of the frame shows sky or ceiling; outdoors, sky in the frame also upsets exposure.
Blind spots don’t only form under the camera. Columns, racks, stacking areas, door leaves and parked vehicles also create shadow areas. Drawing each camera’s view triangle and the shadows of obstacles on the plan reveals overlooked zones. Five minutes on paper, an incident on site. The common solution is to place two cameras crosswise so each sees the area under the other: if the cameras at the two ends of a corridor face each other, each covers the other’s blind zone and also records any interference with the opposite camera.
Recommendation table by area type
The values aren’t a standard; they’re typical ranges used as a starting point on site. Verify each point against the target image level and the angle calculation above.
| Field | Typical height | Purpose | Watch out for |
|---|---|---|---|
| Building entrance, turnstile | 2,2–3 m | Face identification | Small vertical angle, backlight |
| Corridor | 2,5–3 m | Recognition, movement tracking | On the corridor axis, cross coverage |
| Checkout, cashier | Ceiling | Transactions and cash movement | A separate camera for faces |
| Indoor car park | 2,5–3,5 m | Vehicle and person observation | Column shadows, headlight glare |
| Outdoor car park, yard | 4–6 m | General observation | Blind zone, pole vibration |
| Perimeter | 4–6 m | Detection, observation | Vegetation and fence shadows |
In special layouts such as warehouse rack aisles and retail sales floors, these general rules combine with sector-specific constraints; rack aisle geometry is covered in the warehouse and logistics camera design article. The effect of camera angle on video analytics accuracy, rather than on a human viewer, is the subject of the camera angle and analytics accuracy article; analytics algorithms may call for different angle preferences than a human viewer.
Balancing vandalism and access
The low mounting needed for identification exposes the camera to interference. You resolve this contradiction not by increasing height but with layers of protection:
- Impact resistance: At reachable heights, choose a dome or housing with a high IK rating under IEC 62262 (for example IK10). Details of environmental protection ratings are in the IP and IK ratings for outdoor cameras article.
- Cable protection: Having the cable enter the wall directly behind the housing, or run inside metal conduit, makes it harder to cut. An exposed hanging cable makes the cable the target, not the camera.
- Tamper detection: Connect the detection functions for image covering, redirection and defocusing to the VMS as alarms. Tamper detection that’s on but connected to nothing is useless.
- Mutual view: Every reachable camera should be in the frame of at least one other camera.
- Maintenance access: Mounts above 4 m may need a platform for cleaning and fault response; the maintenance plan and its cost depend on this height decision.
With pole mounting, wind-induced sway increases with height. With a long focal length lens, a small sway turns into noticeable shaking in the image; the pole section and fixing detail are as important as the camera choice.
Post-installation checks
In enterprise camera systems, if the installation team leaves a camera at the level of “there’s an image,” geometric errors surface at the first incident review. By then it’s too late to fix them. After installation, for each camera:
- Walk a test person from the planned target point toward the camera’s blind zone; note at what distance the face enters and leaves the frame.
- Review the recording frame by frame to verify that the expected image level (observation, recognition, identification) is achieved at the target point.
- Mark the actual frame on the plan; mark blind zones and camera overlaps.
- Record the mounting height, tilt and pan angles, lens setting and a reference screenshot. This information is part of the camera inventory (how to keep a camera inventory).
- Keep the reference image for comparison during periodic maintenance; a camera’s direction can drift over time because of wind, impacts or cleaning.
- Recheck critical points at night and at the worst sun hour.
The IEC 62676-4 application guidelines offer a general framework for documenting image targets and camera placement; the standard is available from the IEC publications catalog. To produce a blind spot map across the facility and verify mounting geometry at the project stage, see our camera positioning and blind spot analysis and enterprise IP camera installation services.
Frequently asked questions
How high should I mount the camera?
There’s no single right height. At a door where face identification is the goal, the camera sits low and away from the target; in general area monitoring, height provides coverage and protection against vandalism. Choose the height by calculating it together with the vertical viewing angle at the target and the blind zone under the camera.
Why does a high mount make faces unrecognizable?
The higher the camera and the closer to the target, the more the face is viewed from above. As the vertical angle grows, what’s left in the image instead of a face is the forehead, hair and cap peak. Moving the camera back reduces the angle, but then a narrower lens is needed for the same pixel density.
How do you close the blind spot directly under the camera?
The blind zone comes from the camera’s own geometry and doesn’t disappear completely with tilt. Common solutions are a second camera looking from the opposite direction, placing cameras crosswise so each covers the area under the other, or moving the camera back so the critical point stays outside the blind zone.
How are cameras at reachable heights protected?
An impact-resistant (high IK rating) housing, routing the cable through the housing or a protected conduit, a tamper alarm for covering the image and changing direction, and having the camera in another camera’s frame reduce the risk. Plan these measures together with the mounting design.