Road Safety and Traffic Diagnostics Assisted by Automatic Detection and Classification Techniques

Currently, engineers make video recordings and take georeferenced images in the field and then manually analyze these data in the office. This work includes the identification of vehicles, traffic signs, pavement deficiencies and other key elements for road safety studies.

In this context, CPS has developed the project “Road Safety and Traffic Diagnostics Assisted by Automatic Detection and Classification Techniques“, which aims to reduce the time engineers spend on these tasks through the development of software based on artificial intelligence (AI). Through convolutional neural networks, the aim is to automate the detection of these elements, generating accurate and georeferenced inventories. The software will also learn from the engineer’s decisions and diagnoses, continuously improving its accuracy.

This system will be compatible with various capture devices, such as traditional video cameras or 360º cameras with GPS, and will offer interoperability with open source GIS tools, such as QGIS, and BIM systems. This advance will allow engineers to focus on solving complex problems, maximizing the efficiency and quality of traffic and road safety studies.

This project, carried out in the framework of the “SME Innovation Program (INNOVA-CV) 2023 – Innovation in ICTs”, has been financed by the Valencian Institute for Business Competitiveness (IVACE) and co-financed by the European Union.