LiDAR Safety Monitoring: Catching What Cameras Miss
2026-10-03 · data as of 2026-10-12
LiDAR measures distance and height directly by timing returning laser pulses, so it is unaffected by changing light and reports people and vehicles in 3D. Assistive monitoring, however, must be designed differently from a certified safety function.
Have you ever watched a forklift and a worker cross the same aisle at a shift change? Two seconds decide whether that is a near miss or nothing at all.
Teams add cameras to protect those two seconds, but backlight, dust and night lighting often make the footage useless. That is why more sites now pair cameras with LiDAR.
When the light changes, cameras wobble
A camera ultimately reads patterns of brightness. Backlight from a window or a haze of dust blurs the boundary between a person and the background, and detection rates drop noticeably. Turn on lights at night and new shadows appear.
LiDAR looks at distance, not brightness. Even with lights off and a little dust in the air, it returns a person-s position as coordinates. That is why it holds up in dark warehouses and dusty yards.
Is weather a weakness, or a misconception?
People often conclude that LiDAR is bad in rain and fog. That is only half true. A 2023 study by the University of Minnesota and the state DOT found that in fog, LiDAR actually sees farther than cameras or the human eye.
The limits are real, though. A study measured on actual roads (MDPI Sensors, 2023) found that in thick fog with visibility under 50 m, the number of points dropped by up to 59% and intensity by up to 71%. Finland-s VTT reported a 25% reduction in detection distance in fog and snow. Heavy snowfall, or ice on the lens, is genuinely hard for LiDAR too.
Fixed industrial installations, however, face different conditions from self-driving cars. Ranges are short, the unit is often under a roof, and vehicles move slowly — so the same sensor performs far more stably.
Monitoring and safety stops are different problems
One misconception is worth naming: that installing LiDAR is itself a safety certification. It is not.
Protecting people is the job of a separately certified device. A safety laser scanner for an AGV must hold ratings such as IEC 61496 Type 3, ISO 13849 PLd Cat 3 and IEC 61508 SIL 2, and IEC 61496-3 defines its target as objects from 30 to 200 mm. A general-purpose 3D LiDAR, by contrast, excels at watching a wide area and warning early.
They are not competitors; they divide the work. Seeing broadly and warning early is a different requirement from stopping reliably when a person enters a hazard zone.
Before asking how many sensors to install, ask what you are merely assisting and what you are delegating to a certified device. Once that boundary is set, the design gets much simpler.
As long as people and machines share a floor, this question stays. It is worth checking whether your current setup genuinely protects those two seconds.
Source page
https://www.melissasoft.com/en/detect