This is horizontal geometry: walls, tall shelves and the van are treated as opaque. Height and vertical tilt are not considered here; low furniture is decoration. Area is sampled on a 0.25 m grid and does not indicate recording usability.
02 · Height, tilt and detail
Ideal image detail
Pixels across the object
Downward angle to object
Visible ground on the axis
—
Object and resolution
Rectilinear optics and 16:9: VFOV = 2·atan(tan(HFOV/2)/(16/9)). Density = P/(2·Z·tan(HFOV/2)), where Z is object depth along the optical axis. Use your camera resolution and FOV; 16 cm is an example face width. Lighting, face orientation, blur and compression are not modelled, so pixel count does not guarantee identification. Axis · pixel density, IEC 2025 ↗
03 · Light and the recorded image
—
The 0–100% levels are relative, not lux or a sensor model. The image explains contrast and reflections; test WDR and illumination on your camera in daylight and at night. Axis · lighting ↗
Six more comparisons: what the recording retains
× Overview only✓ Overview + face
Recorded: the wide view shows the event; a separate close view preserves facial detail.
01
Separate overview from identification
One wide-angle camera shows the event; another captures a face or number plate at the point of passage.
× Steep angle✓ Face in view
Camera angle and the recorded faceOn the left a high camera looks steeply down and records the top of the head. On the right a more suitable viewpoint shows the face. Recorded frames appear below; no universal mounting height is specified.Recorded: the top of the head from a steep viewpoint; the face from a suitable entrance angle.
02
Do not capture only the top of a head
The steeper the angle from above, the fewer useful facial features remain. Choose a passage point and test the angle with a person.
× Light behind✓ Light on face
Backlight and a well-lit faceOn the left a bright window behind the person turns the face into a silhouette. On the right light from the camera side illuminates the person. Recorded faces are compared below.Recorded: a dark silhouette against a window; facial detail with light from the front.
03
Light matters more than megapixels
A window, headlights or the sun behind a subject turns a face into a silhouette. Change the direction or add even lighting.
× Nearby overhang✓ IR without glare
Infrared reflection from an overhangOn the left some infrared illumination hits a nearby overhang and reflects into the lens, washing out the recording. On the right a clear illumination field produces a clearer nighttime recording.Recorded: a bright IR reflection; a clearer image after redirecting the illumination.
04
Keep IR away from reflections
A wall, glass, dirty dome, spider web or nearby branch can wash out the night image. Remove obstructions from the near field.
× Blind area✓ Two viewpoints
A blind area behind a wall and a second cameraThe floorplan shows a wall between a camera and a person. On the left the person in the blind area is missing from the recording. On the right a second camera sees and records that passage.Recorded: the person behind the wall is missing from the first view and visible in the second.
05
Close blind areas
A wall, building corner or tall furniture can block the view. Give a critical passage its own viewpoint and check continuity between recordings.
× Daytime only✓ Day + night
Checking daytime and nighttime recordingsBoth scenes show nighttime. On the left the system was checked only in daylight; without suitable light the person and face are lost in the nighttime recording. On the right nighttime playback and lighting were checked, keeping the face visible.Recorded: checking nighttime playback reveals darkness, noise and missing detail.
06
Accept the system based on recordings
Check daylight, night, backlight, motion and rain. Live view does not show what will actually remain in the archive.
These diagrams show basic logic, not a finished design. Consider the monitoring objective, privacy and local recording requirements.