The fully automated system is trained using Artificial Intelligence-powered object identification to recognise street signs through Google Street View images.
The results show the system detects road signs with nearly 96 per cent accuracy, identifies their type with nearly 98 per cent accuracy, and can record their precise location from 2D images.
Study lead author and RMIT University Geospatial Science Honours student, Andrew Campbell said the proof-of-concept model was trained to see ‘stop’ and ‘give way’ signs, but could be trained to identify many other inputs.
He said the system was easily scalable for local governments and traffic authorities.
Mr. Campbell said large amounts of time and money are spent recording the geo-location of traffic infrastructure manually, using a process that exposes workers to unnecessary traffic risk.
“Councils have requirements to monitor this infrastructure but currently have no cheap or efficient way to do so,” Mr. Campbell said.
“By using free and open source tools, we’ve now developed a fully automated system for doing that job, and doing it more accurately.”
The RMIT research team found mandatory GPS location data in existing databases for street signs was often inaccurate.
“Our system, once set up, can be used by any spatial analyst – you just tell the system which area you want to monitor and it looks after it for you,” Mr. Campbell said.
RMIT geospatial scientist and project co-lead Dr. Chayn Sun said some councils already attach cameras onto waste trucks to gather street footage, demonstrating how valuable visual data is becoming.
“Where footage is already being gathered, our research can provide councils with an economical tool to drive insights and data from an existing resource,” Mrs. Sun said.
The project was co-led by Mrs. Sun and fellow RMIT geospatial scientist Dr. Alan Both, from the university’s Centre for Urban Research.