TTS Research - Crack detection on local roads. Phase 1

The measurement of cracking forms an important component of TTS surveys, and also of current visual condition surveys such as CVI and DVI. The presence of cracking is often a trigger for maintenance. Consequently it is important that cracking is measured and reported as accurately as possible in SCANNER surveys. However, experience with the automated measurements of cracking provided by TTS survey vehicles has shown that the measurement of this defect is in need of improvement. It is likely that the move to lower classes of road, which present greater challenges SCANNER system, is likely to emphasise any shortcomings. This report presents the results of the first phase of a project aimed at developing the measurement of cracking on local roads, and hence defining a specification for such systems for the SCANNER survey commencing in 2007. This phase has aimed to identify the basic requirements for the measurement of cracking, from the viewpoint of the end user (engineer), then to identify and assess currently available crack measurement systems that offer potential for enhancement to meet the requirements of SCANNER surveys on local roads. Following the consideration of equipment offered worldwide, the review short-listed five systems for testing, on the basis of capability demonstrated to date, potential capability, and practicality for implementation in the UK within the next few years. The system operators were: TRL (operating HARRIS on behalf of the Highways Agency); WayLink; DCL; WDM; Jacobs-Babtie. The systems have been assessed on local road sites located in Leicestershire. All of the systems are able to identify cracks, with varying levels of accuracy. The work did not identify a system that clearly outperforms the others. All of the systems tended to have problems with common non-cracking features present on road surface, including joints, patches, fretting, ironwork, high friction surfaces, road edges and white lines (to a lesser extent). All of the systems are capable of correctly ignoring these features in some locations, but report false positive cracking in other locations. All of the systems would have difficulties meeting the current requirements for acceptance on the test sites, primarily as a result of the effect of false positive crack detection on the statistical analysis. (A)