© Tutti i diritti riservati
ARGUS - AUTOMATED RECOGNITION OF GUARDRAIL SYSTEMS
Every year, road accidents claim thousands of lives across Europe. Some of the most tragic events are linked to outdated or non-code-conforming bridge barriers. But what if we could automatically detect these barriers and help road authorities prioritise interventions?
Let’s recall two dramatic accidents. The Acqualonga Viaduct tragedy in 2013, where a bus with 48 people on board crashed, resulting in 40 fatalities. And the Mestre accident of October 2023, with 22 victims and 14 injured.
These incidents raise a crucial question: How can infrastructure managers not be aware of the condition of their bridge barriers?
The answer lies in the lack of simple and scalable tools. This is where the ARGUS system comes in –
Automated Recognition of GUardrail Systems. ARGUS uses artificial intelligence and open data to analyse bridge guardrails on a large scale. It relies on YOLO – “You Only Look Once” – a fast and powerful object detection algorithm that identifies barriers directly from images. The workflow is fully automated.
First, images are collected using services like Google Street View. Then, the trained AI model detects and classifies the barriers, drawing bounding boxes and assigning confidence levels. The result is a regional map of bridge guardrails, identifying where maintenance or replacement is urgently needed. This system offers several strengths.
It is fast, scalable, and usable in real time.
You can watch the video https://www.youtube.com/watch?v=R4iVQZKm49Y
