Road safety 3D modeling

Road intersections survey and modeling for the design and evaluation of safety improvement measures / High risk locations all over Greece


The project was assigned by the Maintenance Department of the Greek Ministry of Infrastructures, Transports and Networks in order to improve safety for high accident risk locations (mainly road junctions) all over the Greek road network and, in particular, to design and evaluate through video simulations the appropriate improvement interventions for these locations.

Project tasks

  • Selection of “high accident occurrence – reduced safety” locations by the Maintenance Department.
  • Laser scanning survey, using the Optech ILRIS 3D Laser Scanner, of the road intersection area for each location. Sufficient road length had to be covered in order to create driving video simulations, at least 200 – 300 m for each intersecting road. Required resolution was about 5-10 cm, with an accuracy of 5 cm. Scanner was placed on the roof of a vehicle or lifted by a lifting device. Approximately 10 scanning positions were required per location, with a distance between scans 50 – 80 m and 20% (maximum) pan-tilt base overlap was used.
  • Alignment and georeference of the pointclouds to create a single colored and georeferenced point cloud for the surveyed area. Georeference was performed using the scanning stations as well as conical targets, surveyed with GPS from state geodetic control points.
  • Vector drawings (feature collection) and orthophotos (directly from colored scanned points) were created as a background for design purposes.
  • Design of safety improvement interventions: additional lanes (including acceleration – deceleration lanes), vertical and horizontal signage, changes in traffic regulation, closure of secondary roads, etc. Design was performed in vector format and new road elements were modeled in 3D (as polygonal models).
  • Vertical signage modeling (necessary for proper display of traffic signs in 3D video simulations) was performed by creating a small VRML model for each traffic sign. Traffic sign pictures were overlaid on circular, triangular, etc, 3D surfaces and exported as VRML models. Standard traffic sign pictures were available, while digital pictures of non-standard signs (e.g. showing directions) were taken and photos were ortho-rectified based on the shape of the sign. Finally, new (according to the design) signage pictures were created using image processing software.
  • Editing – cleaning of the initial pointcloud was performed and polygonal models of new road elements, as well as signage VRML models, were imported in a single environment, to create the “as designed” total 3D model.
  • Video fly – through and driving simulations were created. Driving simulations (from all possible origins to all possible destinations) were generated by applying vehicle trajectories and calculating distances and video frame intervals to simulate movement with variable speed. The same video frame sequences were applied to “before” and “after” models to create comparative results.
  • Video simulations were evaluated by the staff of the Maintenance Department and the Local Authorities in order to approve the designed measures for construction or forward the design for further revision and improvement.

Results – Conclusions

Laser scanning has proven to be not only a fast data capture tool to provide accurate and complete survey background data for road design purposes, but also an easy and efficient way of creating “before / existing” and “after / as designed” colored video fly-through and driving simulations to easily compare and evaluate the efficiency of designed – proposed interventions, before their actual implementation.

Related Videos