Tallinn's buses help identify missing and damaged traffic signs

Researchers at Tallinn University of Technology have installed AI-based camera systems on three Tallinn city buses to automatically detect damaged or poorly visible traffic signs among the city's approximately 50,000 signs, with data processed onboard and sensitive footage blurred. The project, part of the international TayuTech initiative led by the FinEst Smart City Center of Excellence, aims to speed up inspections that currently take weeks and could expand to detecting road damage or streetlight failures. Data collection runs until the end of November, after which the system's accuracy will be assessed.

Tallinn's buses help identify missing and damaged traffic signs To map the condition of traffic signs, researchers at Tallinn University of Technology have installed sensors on the front of three Tallinn city buses that collect visual information while driving. Researchers at Tallinn University of Technology installed sensors and AI‑based camera systems on Tallinn city buses to detect poorly visible or damaged traffic signs. The goal of the project is to obtain a better overview of the condition of traffic signs and thereby improve traffic safety. Tallinn has approximately 50,000 traffic signs, and regularly monitoring their condition is a resource‑intensive task. Currently, traffic signs are inspected by people, and identifying deficiencies can take weeks. Now researchers are testing how this task could be automated and made faster. City buses as information gatherers To do this, researchers at Tallinn University of Technology installed sensors on the front of three Tallinn city buses, which collect visual information about traffic signs while driving. Although using buses makes it possible to cover a large portion of city streets, they still do not travel everywhere where information would be needed, noted senior researcher and project lead Uljana Reinsalu from Tallinn University of Technology. For this reason, the researchers are considering alternative solutions. "We need to find other types of vehicles that allow better coverage. One option is to use garbage trucks," Reinsalu explained. According to the researcher, the data are processed with artificial intelligence already inside the vehicle, and only the necessary results are sent to the city. Since the system uses video footage that inevitably captures other road users, the researchers consulted the university's research ethics experts and the Data Protection Inspectorate. "Our current algorithms detect both people and vehicle license plates, and we automatically blur them," Reinsalu added. Possible other applications In addition to mapping damaged traffic signs, the same platform could in the future be used to detect other issues in the urban environment, such as road surface damage or street‑lighting failures. According to Reinsalu, information about air quality is already being collected today. The researchers want to demonstrate how solutions based on mobile sensors and artificial intelligence can make city management more efficient, data‑driven, and sustainable. According to Reinsalu, information related to the urban environment is currently scattered across different places. "We are trying to conduct monitoring in a way that is thorough and structured, so that city employees can access the information and it helps them find solutions and make decisions faster and better," she explained. Researchers will collect data from city streets until the end of November, after which they will assess how accurately the system can detect traffic signs and how well the solution fits Tallinn's needs. This is part of the international project TayuTech , led by the FinEst Smart City Center of Excellence. In Tallinn, the focus is on monitoring the condition of traffic signs; in Dublin, Ireland, air quality is measured; and in Pula, Croatia, road surface damage is detected. -- Editor: Argo Ideon