Automated 3D Mapping, localization and pavement inspection with low cost RGB-D cameras and IMUs
Accepted version
Peer-reviewed
Repository URI
Repository DOI
Change log
Authors
Abstract
Roads play a critical role in a country’s infrastructure by facilitating the transportation of people and goods. The Department for Transport in Great Britain revealed 297.6 billion vehicle miles on Great Britain’s roads in 2021. As traffic volume and road age increase, a range of irregularities can emerge on the road surface. Although numerous studies have already focused on the state of road pavements, there is still potential for the use of low-cost technology for the swift inspection of road pavements. This study aims to bring low-cost technology to help road inspection to be more efficient and accurate. To this end, a fast realtime condition monitoring approach is proposed for road pavements using an RGB (Zed 2i) stereo camera with IMUs mounted on a car. The algorithm used for road pavement reconstruction is based on Structure from Motion (SfM) and requires no calibration of the sensor data. The experiment uses two types of roads to compare the pavement reconstruction and IMU data. The findings indicate that the method effectively reproduces pavement distress, promoting a preliminary approach to enhance pavement management.