The Highways Agency conducts routine automated surveys of trunk road pavement surface condition under the TRACS survey, and of structural condition under the TRASS survey. These surveys assist the Agency in establishing the condition of the network, prioritising reconstruction and repair work and valuing its road assets. However, the network includes assets other than pavements – such as bridges, signs, gantries, lighting, vehicle restraint systems, embankments and many types of roadside furniture – none of which are assessed using current automated routine surveys. LIDAR (Light Detection and Ranging), which uses laser sensors to measure distances from the sensor head, has the potential to deliver measurements of objects much further from the survey vehicle than TRACS surveys. Although the accuracy of LIDAR sensors is lower than TRACS sensors, it is well within the accuracy required for measuring many roadside infrastructure assets. This report describes an investigation into the use of LIDAR for the measurement and assessment of road assets. The work, which builds on previous research carried out for the Agency, focusses on the automation of asset assessment, describing the development and application of algorithms for the location and measurement of roadside barriers, bridges and gantries. The algorithms draw upon a method called LIDAR slicing to allow the large LIDAR data set to be broken down into manageable lengths whilst still containing the information required for automated analysis. An algorithm to assess barriers is shown to automatically identify the presence of barriers (either steel or concrete) and estimate their heights. Testing using data collected on the M25 shows it to be capable of identifying the heights of 80% of barriers to an accuracy of ±5cm. There is a reasonably low level of noise in the data, but errors occur where bridges or vegetation lie between the survey vehicle and the barrier. An algorithm to locate and assess bridges has also been developed and tested on structures (bridges, foot bridges and gantries) on a section of the M25. It is shown to identify 25 of the 26 structures. Furthermore, the algorithm correctly discerns the 8 road bridges from the 26 structures. Automated bridge heights show good agreement with those obtained from manual assessment. It is suggested that the algorithms could be applied on the network, but it would be preferable to fine-tune their using network tests before implementation. Further work could also consider how the data could be used to assist the road operator in managing the asset. With the size of the HA’s network and its many thousands of asset inventory items there would be significant advantage in time, cost and objectivity if the use of automation could be expanded. This work has shown that LIDAR could offer the potential to automate the identification and measurement of these assets. However, the work has been limited to two distinct types of asset, representative of the types of asset that would be most straightforward to extract from LIDAR data. There is potential to expand the automated application to other assets, although this would require the development of further algorithms. The required LIDAR data could potentially be provided through the addition of LIDAR to existing TRACS surveys. However, before considering expansion of the LIDAR methodology there is a need to consider how such data would fit into the Agency’s current Asset Management practice, in terms of the need for the data, the required coverage and accuracy and the required frequency of updates. This would assist in establishing a case for further development and implementation. (Author/publisher)
Samenvatting