Near-Miss Detection
See the risks before they become accidents
Accident records are an important source of data, but they only tell you what has already happened. They say nothing about the risky situations that occur every day but do not result in a collision. Relying only on crash statistics means waiting until people get hurt before taking action.
Flow Analytics by AGC changes that. With LiDAR-based, privacy-first analytics, we detect near-miss interactions between vehicles, cyclists and pedestrians in real time. This gives cities and transport authorities a clear picture of where the dangers are, long before they appear in official crash reports.
Why Near-Miss Detection Matters
Near-miss detection provides an early warning system for urban safety. It allows decision-makers to intervene before accidents happen, saving lives and reducing costs.
How It Works
Our LiDAR sensors track every road user at centimeter-level accuracy and reconstruct their paths through the intersection or street segment. Flow Analytics algorithms then analyse these trajectories to detect conflicts using proven safety metrics such as:
- Time to Collision (TTC) – the time remaining before two users would collide if they continued at the same speed and direction.
- Post Encroachment Time (PET) – the gap between one user leaving a conflict point and another entering it.
- Speed and trajectory overlap – indicators of high-risk behaviours such as sharp braking, swerving or failure to yield.
The result is a detailed record of where and when near misses occur, which modes are involved, and how severe the risks are.
From Insights to Action
Near-miss analytics turn invisible risks into clear priorities. With Flow Analytics by AGC, you can:
- Map hotspots where conflicts are concentrated.
- Rank intersections or crossings by risk exposure.
- Compare before and after conditions when signal timings or layouts change.
- Support Vision Zero and Safe System policies with measurable evidence.
Recent Near-Miss Detection Implementation
Flow Analytics by AGC introduced its near-miss detection in Oulu, using LiDAR analytics to uncover conflict zones where vehicles, cyclists and pedestrians nearly collided. The insights helped city planners identify and prioritise public safety interventions.
Why Near-Miss Detection from Flow Analytics by AGC?
Privacy-first
Our solution never captures faces, license plates or personal data. It relies on anonymous 3D LiDAR points, making it fully GDPR-compliant and safe for public use.
Accurate and detailed
Movements are tracked down to the centimeter, enabling precise counts, classifications and trajectory analysis that manual methods and cameras cannot match.
Proven in real-world conditions
Designed for complex urban settings with heavy traffic, poor lighting or weather challenges, ensuring consistent results where other systems fail.
Actionable
Instead of raw feeds that need extra processing, we deliver insights in dashboards, reports and GIS formats that decision-makers can apply directly to planning, design and operations.