Talk by Kris Vanherle sharing his story of building Telraam, a large crowd-sourced platform of traffic counting devices that empowers a growing international community of citizens to influence local mobility policy.
4. Telraam, a sensor in the window
A sensor in the window with a camera that counts
continuously (in daylight) and is able to
distinguish multiple modes, direction and speed
(*):
● pedestrians
● bikes
● cars *
● motorcycles
● vans *
● trucks *
● buses
6. Components of the Telraam S2
Purpose-built Telraam
Custom PCB
Dedicated AI chip
IoT framework
USB C-powered
Built-in camera
Wide angle camera
Auto region of interest selection
LCD display
Interaction with device
Instant visualisation of
counting data
AWS server
Classification
Data aggregation/reports
API
Data connection
AI counts on device
No images saved
No wifi needed
7. Uses of the Telraam S2
AI generated counts
Ability to improve through
machine learning
Easy to install
10’ installation process
Plug & play
Compact
All-in one & elegant design
LCD interface
More traffic data
Expanding to more modes
15’ typical traffic
Night counts *
Counting motorised
vehicles in darkness
* night counting functionality in development & testing
10. ✅ Clear view (= no flyscreen, no shutters, no big trees, …)
✅ 1st/2nd floor
✅ Not too far from the street (max. +-15m)
✅ Camera aimed at the middle of the road
✅ Not near traffic lights or junctions
✅ No movement not related to traffic
✅ Availability of power / plug
Around 65% of the houses are suited to install a Telraam
Technical criteria for Telraam installation
11. Privacy:
● no ANPR
● no face recognition
● no route tracking
● no video images
● image processing ‘on the edge’
24. Traffic speed reduction after Telraam
(speed below and above 30 km/h %)
Impact of:
(a) placement of digital speed sign
(b) placement of Berlin cushion on average speed in zone 30.