Road safety could become automated

Researchers aim to adapt ‘image pattern recognition’ software to enable automated systems for the analysis of road safety and road conditions.

Their project aims to build on data already collected through vehicle-mounted video and mobile laser scanning.

Successful outcomes could find automated solutions for monitoring required by national and international safety and risk assessment models.

“Much of the current data is inaccurate due to changing conditions and it is processed manually,” said the project Chief Investigator, CQUniversity’s Professor Brijesh Verma.

Professor Verma is working on the project with partners from the Australian Road Research Board (Dr Joseph Affum) and the Department of Transport and Main Roads (Dr David Stockwell).

“The QLD Department of Transport and Main Roads (DTMR) collects a variety of data to record and ascertain the status of road safety and condition, and uses these data extensively both for daily operations such as wide load permits and road maintenance and for guiding and justifying capital expenditures,” Professor Verma says.

“Vehicle mounted video (DVR) is collected over every state road annually, and Mobile Laser Scanning (MLS) data is collected periodically.

“Automation of the extraction of road attributes from DVR video using advanced image analysis, and cross-validation with other data sources such as MLS, has the potential to provide a range of value-added products consistently and inexpensively: road condition (e.g. deflection, cracking, rutting), road safety (e.g. selected AusRAP road safety rating attributes and iRAP International Road Assessment Program star rating compliance), environmental (e.g. fire risk and vegetation encroachment), and improved obstacle clearance estimates (e.g. overhead wires, roadside hazards, heavy vehicle widths).

“The aim of this project is to develop and evaluate deep learning neural network based methods for extracting road attributes, especially those required by AusRAP/iRAP and Australian National Risk Assessment (ANRAM) models.”

Professor Verma says initial stages of the project would inform a cost/benefit analysis of options to progress the AusRAP road data collection process for implementation into software.

This research project is funded by the Australian Road Research Board (ARRB) and research is conducted within CQUniversity’s Centre for Intelligent Systems.

Source: CQU

Most Popular

To Top