Highway as-built surveying

As-built survey and 3D modeling of Highways in Greece: Korinthos – Tripoli highway (static TLS), Elefsina – Korinthos highway (mobile TLS)

Purpose

Concession Self Financing Projects have been, during the last decade, a common practice for the construction of road transport networks. The basic concept is that a large private J/V undertakes the construction of a new highway section and in the same time takes the responsibility for the maintenance (and improvement) of an existing highway section, from which it collects the toll fees, in order to finance the whole project. After completing the project, the J/V has its full exploitation for a certain number of years, according to the concession contract.

An obvious need for detail surveys of the existing highway sections arises from the whole process. These surveys usually require:

  • Detail “as-built” survey of all highway features (pavement, structures, slopes, signage, poles, etc).
  • Efficient archiving of “as-built” situation for future reference.
  • Positional accuracy: 2-3 cm.
  • Elevation accuracy: 1-2 cm.
  • 3D model (TIN) for highway reconstruction design.
  • Background survey maps (scale 1:500).
  • No significant traffic closure or delay.
  • Efficient safety plan.
  • Permits from local traffic authorities.

Applying TLS methodology

Terrestrial Laser Scanning techniques have been applied in two cases of existing highways (dual carriageway, 2-3 lanes & shoulder), using two different approaches:

Korinthos – Tripoli highway (length 80 km, J/V MOREAS) was surveyed in 2006-2007 using a static (scan & go) approach with an Optech ILRIS 36D Laser Scanner.
Elefsina – Korinthos highway (length 60 km, J/V APION KLEOS) was surveyed in 2008 using a mobile approach with the newest Optech LYNX Mobile Mapper.

Project Tasks

a. Korinthos – Tripoli: Field work tasks and parameters

  • Establishment of geodetic infrastructure networks (triangulation, leveling, polygonometry), also necessary for construction.
  • Static (stop & scan) laser scanning with Optech ILRIS36D.
  • Scanner carried by a vehicle moving or standing always on the shoulder lane, protected by a traffic regulation trailing vehicle.
  • Scanning from both sides of the highway, distance between scanning positions 50-80m.
    1100 total scanning stations, 120 working days for 80 km of highway.
  • Critical issue for horizontal objects: lifting the scanner (better scanning angle, improved object visibility, lower scanning resolution and / or fewer scanning positions required).
  • Lifting device used: Genie Super Hoist (5.6m, 113 kg capacity, CO2).
  • Custom modifications: Trailer integration, 5/8 bolt, longer ethernet and power cables, stabilizers, fuel generator & UPS, etc.
  • Scanning resolution: 55mm @ 25m horizontal / 20mm @ 25m vertical.
  • Pan-tilt base overlap set to maximum (20% overlap, 15 frames/3600).
  • Primary georeferencing: with conic targets (standard traffic cones: easy to install, measure and model), 1 cone (anchor point) per scan position required for sequential georeferencing.

b. Elefsina – Korinthos: Field work tasks and parameters

  • Establishment of geodetic infrastructure networks (triangulation, leveling, polygonometry), also necessary for construction.
  • Mobile laser scanning with Optech LYNX Mobile Mapper (collaboration with SINECO).
    Sensors – GPS/IMU carried by a vehicle moving at 50 km/h on the shoulder and left lane, protected by traffic regulation vehicles.
  • 2 passes for each carriageway (shoulder lane – left lane) for better data quality.
    240 km total scanning distance, 1 working day for 60 km of highway.
  • Base GPS station support (6 base stations on known points).
  • Measurement of positional Ground Control Points (natural targets identifiable on pointcloud).
  • Basic data processing / alignment and delivery of georeferenced pointclouds in 500 m segments for each carriageway.
  • Conversions between global (WGS84/UTM/zone 34) and local (CGRS87) geodetic reference systems.
  • Positional GCP alignment for groups of 3-5 segments of 500 m (typical target registration accuracy < 3cm).

c. Common post-processing tasks

  • Georeferencing refinement for elevations: using additional points measured on both edges of each carriageway every 50-80m (typical elevation alignment accuracy < 1 cm).
  • Feature collection from pointclouds.
  • 3D Modeling (TIN) from features and Survey Maps (scale 1:500) generation.
  • Archiving for future reference: Pointclouds segmented per km.

 

This slideshow requires JavaScript.

Results – Conclusions

While the newest mobile TLS approach using the LYNX Mobile Mapper is obviously the method of choice, the static approach with the ILRIS still has some advantages and can be applied at least for smaller road sections, taking into account also the cost of the two systems. The comparison conclusions between the two methods, in terms of data quality, accuracy and productivity are presented below:

Data Quality

LYNX:

  • Uniform resolution homogeneous pointclouds.
  • No unnecessary overlaps.
  • Less noise from passing traffic.
  • Better object coverage with 2 sensors.

ILRIS:

  • Better detail for close objects.
  • Better viewing angle when lifted.
  • Produces organized pointclouds (with normal vectors).

Accuracy

LYNX:

  • No errors from overlapping frame ICP alignment.
  • No errors from sequential scan positions ICP alignment.
  • Good relative accuracy for segments of 500 m.

ILRIS:

  • No errors from GPS outage or poor satellite conditions.
  • No errors from attitude compensation.
  • Excellent relative accuracy for each frame.
  • Lifting device can lower accuracy with bad weather conditions.

Productivity

LYNX:

  • Field works: Dramatically faster (1 day vs months) and safer.
  • Faster alignment and georeferencing of datasets.
  • Significantly faster and easier noise cleaning.
  • Automated feature extraction tools work better with uniform density homogeneous pointclouds.

ILRIS:

  • Easier manual feature collection with shaded organized pointclouds.
  • Advanced filtering techniques work only with organized pointclouds.
  • Better level of detail for close objects (resolution – viewing angle).

Related Videos