track service writing

Visual AI for transport networks monitoring

TraCK, the Transport Camera Kit, is a collection of advanced visual AI analytics to perform road assets inventory while monitoring their quality. Feed your day-by-day operations and predictive maintenance with actual data from the field, captured with a single pass.

Discover more

Turn connected vehicle cameras into active road sentinels

Add unprecedented capabilities to your cameras in 3 simple steps:

1
INSTALL

TraCK analytics can run both on vehicle PCs to process images in real time and on external server for post processing mode.

2
ANALYSE

TraCK analytics create a database of road assets and defects, assigning to each detected feature its position.

3
REACT

Receive real time notifications directly from vehicles on the road, or browse the most relevant data collected by TraCK when they come back.

futuristic image of a person on the computer watching a city from it

How it works?

TraCK analytics work with images captured by vehicle cameras to automatically generate information and reports on the road pavement and road assets conditions. TraCK analytics run both in real-time on a connected vehicle PC equipped with GPS, to perform the most critical analyses, and in post-processing when the vehicle returns to base, to extract information that doesn't need to be transmitted immediately.

symbol of a down arrow

Key Benefits

TraCK analytics return geolocated information on many aspects of the road infrastructure, like road pavement and road signs conditions, street lamp lights status or vegetation encroachment. You may equip your own vehicles with cameras and an automotive grade PC for continuous and frequent monitoring, or you may rely on our camera kit to use any vehicle for occasional surveys. No matter what’s your choice, TraCK analytics will always capture images and provide data at cruising speed, up to 100 km/h.

Request information

Key Features

Data returned by TraCK analytics are structured in JSON format, containing information on the vehicle position, a timestamp, and descriptions of the identified features. Depending on the analytics, these can be coupled with supporting visuals like graphs, camera snapshots or time lapse videos. TraCK data are immediately ready to be used in control room dashboards, managed with a GIS or exchanged with any third-party decision support system.

Analytics

image of a road with road signs

Road signs

Visual AI for road signs inventory and monitoring, on the go

This function analyses images captured at high frame rate from a vehicle camera to identify and geolocate the road signs along the way. The type and position of each road sign is stored in a database, to identify missing elements and perform automated analyses on their conditions.

Street lamp lights

Visual AI for periodic inspection of street lamp lights status and conditions

TraCK adapts camera parameters to enhance the visibility of different features, like street lamp lights in road and urban environments. This way our visual AI can count, position and report the number of lights that are correctly working.

image from the top of a rail

Vegetation

Visual AI for inspecting the interference of vegetation along transport networks

Plant growth is among the major concerns in  infrastructure maintenance. TraCK identifies herbs, shrubs or branches that may interfere with ordinary operations, and provides geolocated data to plan cleaning interventions and keep a proper safety margin.

image of potholes in the street

Road pavement

Visual AI for mapping road surface defects like cracks, potholes and discontinuities

When it comes to identify potholes, cracks and discontinuities on road pavement, cameras and visual AI come into play. Whether they are processed in real time or not, the generated data can be displayed on maps or used to calculate road pavement quality indices, thus streamlining maintenance planning.

Powered by

White DECK logo on transparent background

TraCK is powered by DECK, the service-oriented software platform developed by WaterView to run visual AI on dedicated servers or edge AI units. DECK works with virtually any camera, and exchanges data with third-party solutions by using secure standard IoT communication protocols. DECK is a Swiss knife offering several tools like universal grabber for camera interfacing, native image anonymization, advanced load balancing, services for remote maintenance and diagnostics, as well as versatile data exchange via MQTT, Kafka, email, Telegram and more.

Discover DECK