Home / Cities and public spaces / Pedestrian safety analytics

Near-miss counts show which crossings are dangerous before someone is hurt.

Road safety teams usually learn that a crossing is dangerous from its injury record. ZoneIntel AI counts the near misses between vehicles and people at each crossing, so the worst ones can be fixed first.

ZoneIntel AI Road safetyDistrict 2 ยท 38 near misses
Paths cross 1.2 s apartVehicle pathPerson pathCrossing X-08, Thu 17:52

Near misses this month, by crossing

  • X-08 Market Road14
  • X-03 School Lane11
  • X-11 Station exit7
  • X-05 Park gate4
  • X-02 Clinic2

A near miss is counted when a vehicle and a person pass the same point less than 1.5 s apart.

People are counted and never identifiedEach near miss keeps a short clip with faces blurred
Simulation. Sample data.

How pedestrian safety analytics works

A near miss is measured from the paths of the vehicle and the person.

  1. Mark the crossingThe crossing and its approach lanes are marked on the camera view.
  2. Paths are followedZoneIntel follows each vehicle and each person through the crossing.
  3. Near misses are countedA near miss is logged when the two pass the same point within the time gap you set.
  4. Crossings are rankedEach crossing is ranked by near misses per thousand people crossing.

What you receive

Road safety teams and town councils get a ranked list to plan works from.

  • A ranked list of crossingsNear misses by crossing, hour and vehicle type.
  • A clip for each near missA few seconds of video, with faces blurred.
  • Before and after figuresThe same count after a new signal, hump or sign.
  • Counts of peoplePeople walking and cycling, by direction and hour.

Questions about pedestrian safety analytics

What counts as a near miss?

A near miss is logged when a vehicle and a person pass the same point within a short time gap. The city sets the gap, and many teams start at 1.5 seconds.

Does it identify drivers or pedestrians?

The analysis works on paths and timing. It reads no plates and keeps no face templates. Clips kept for review show blurred faces and plates.

How is this used by a road safety team?

Teams rank crossings by near misses and plan works at the worst ones. After the change, the same camera shows if near misses fell.

Is it used for enforcement?

Near-miss counts are a planning measure. They show where the road layout or signal timing needs attention. Enforcement stays with the traffic police.

Near misses at one of your crossings

Share one week of footage from one crossing. We send back the near misses by hour, with a clip of each.