SINELEC

Streamlining the control of uncollected automatic tolls thanks to rapid license plate identification

AI & Data Solutions

industry

IT Software

know how

AI & Machine Learning

1 Starting Point

1 Need

Sinelec, a key technology player within the ASTM Group and the world’s second-largest operator of toll highway networks as well as a global player in the design and construction of major infrastructure projects, had the following needs:

Reduction of photo processing times
Speed and efficiency in transit control
Automated analysis of vehicle license plates
Easier and faster classification of vehicle nationalities
A fast but above all reliable image analysis-based system

2 Discovery

2 Direction

Dinova designed and implemented for SINELEC an ALPR (Automatic License Plate Recognition) system integrated with Google Cloud Platform, which allowed for:

Vehicle type classification (including toll-exempt vehicles)
Extraction and optical reading of license plate characters
Classification of license plate nationality using a custom template
Image validation to ensure reliability and speed of analysis

3 How

3 The challenge

The company operates across several Italian highway sections, for a total of over 1,400 km, mainly located in the northwestern regions of the country, such as Lombardy, Piedmont, Aosta Valley, Liguria, and Tuscany.

Every day, in addition to managing operations and traffic, the company is responsible for toll collection and the subsequent verification of unpaid tolls. This process requires analyzing the license plates of vehicles for which an abnormal transaction is detected. Given the high traffic volume, the number of images to be examined is very large.

That’s why Sinelec turned to Dinova to implement a tool capable of speeding up the verification of unpaid automatic tolls — a system, therefore, able to automate the extraction of information from license plates and their validation in shorter times.

To this was added another request from the client: to determine the nationality of the vehicle in order to avoid mistakenly sending Italian payment notices abroad.

4 What

4 Solution by Dinova

Dinova has designed an ALPR (Automatic License Plate Recognition) system to which the license plate nationality recognition functionality has been added..

ALPR is a system that uses a technology called Optical Character Recognition (OCR) to read vehicle license plates under any weather conditions and generate data about their location.

In essence, the implemented system processes images very quickly and analyzes their content by leveraging the capabilities of Google Cloud Platform services, in particular Vertex AI for the Artificial Intelligence components, a Google OCR API for license plate character extraction, and DataFlow for computing power.

The system built for Sinelec is a complex system composed of different modules, each with its own peculiarities and functions:

Vehicle type classification (also detecting toll-exempt vehicles, such as emergency vehicles or law enforcement vehicles) and license plate identification, using Google's Vertex AI;
Identifying and extracting license plate characters. For this, Injenia used a Google API for OCR (optical character recognition);
License plate nationality classification. For this feature, a custom internal model was developed that can assign a license plate's nationality based on its topological characteristics.

5 Why

5 Why Dinova?

AI for reliable and fast analysis of photos and images

After a year of intensive work, the collaboration between Dinova and Sinelec led to the creation of a system with outstanding performance, capable of providing reliable information on the analyzed images and significantly optimizing operator activities.

Thanks to a technology designed with extremely high precision standards, the required data are now available in less than an hour, reducing analysis and evaluation errors. This project demonstrates the value of artificial intelligence at the service of people and businesses, freeing up valuable time to dedicate to activities of greater strategic value.