This innovation activity in the field of digital cities is funded and carried out within EIT Digital, a leading European organization for digital transformation linked to the European Commission, and involves the participation of the Polytechnic University of Milan. It aims to build upon previous research to refine and test in a real-world traffic scenario a fully marketable and unique product that automatically identifies, in real time and with high precision, the type of vehicle traveling on a road and the number of occupants, both in the front and rear seats.
The solution will allow municipal authorities and other transport infrastructure managers, such as road or parking operators, to understand mobility patterns and establish strategies and policies that reduce traffic congestion, prioritize and encourage the use of public transport, high-occupancy vehicles and low-emission vehicles, with the consequent improvement of traffic flow, air quality and noise levels.
The accurate and automatic identification of vehicles and occupants, combined with data processing and analysis, will facilitate a better understanding of traffic, the application of discounts or penalties, variable pricing (for example, in parking lots or tolls), and access restrictions to certain roads, especially in city centers, based on the number of passengers, vehicle type, license plate, etc. It will also help promote the use of public transportation, car-sharing, high-occupancy vehicles, low-emission vehicles, park-and-ride facilities, and other alternatives among citizens.
Currently, implementing these types of measures requires monitoring and deterrent controls by traffic authorities for compliance and the detection of violations, making their widespread adoption complex, ineffective, and unreliable. BeCamGreen aims to finalize the development of an automated, reliable, and reasonably priced commercial product to address a real market need. This solution is increasingly being requested in highway tenders in countries like the United States, facilitating the creation of dedicated lanes for specific vehicles and the development of traffic restriction strategies being implemented by numerous European cities.
Computer vision, deep learning, and multispectral analysis
BeCamGreen will develop an automatic and non-intrusive solution thanks to the application of the latest technologies in big data, artificial vision, deep learning and multispectral analysis.
Indra will work on evolving image processing algorithms for person and facial detection developed in previous R&D projects, such as DAVAO. To achieve greater accuracy, the company will incorporate improved video surveillance equipment and combine these algorithms with new ones developed for real-time image processing. The solution will also include multispectral analysis, which allows for the detection of human skin to avoid false positives and errors, differentiating, for example, a mannequin or other types of simulations. The goal is to incorporate the latest technology, both in hardware and software, to increase system accuracy and reduce investment and operating costs for potential clients.
For its part, the Polytechnic University of Milan will work on developing a big data engine to detect and predict traffic conditions using and integrating real-time information from all types of IoT sensors, social networks, various open data sources, and the project's own vision subsystem. This real-time big data engine will provide valuable information that will help managers make decisions, validate, and improve their mobility management strategies.
