A PRACTICAL SYSTEM FOR CRACK IMAGE COLLECTION AND AUTOMATIC IDENTIFICATION

Recent evidence has demonstrated that the deterioration of thick, well constructed flexible pavements is not structural and that deterioration generally starts at the surface in the form of cracking and rutting. Such behaviour of long life roads emphasises the need for accurate repeatable and economical methods of monitoring and correcting these defects before they affect the structural integrity of the pavement. This paper describes the development of a system for the automatic detection of surface cracking. Automatic analysis depends on obtaining uniformly bright images of high quality. These are collected at survey speeds of up to 80 km/h using novel linescan video cameras operating with artificial illumination. Surveys can thus proceed either by day or night. Purpose built hardware on board the survey vehicle assembles the line images into video frames prior to automatic processing to identify cracks. Firmware is used to carry out part of this processing in real time, the remainder of the processing being carried out off-line on a PC. In such a system the correct application of defect information depends on accurate knowledge of its location. In the UK such referencing is currently provided by accurately measuring the distance travelled with respect to automatically recognised studs on the road surface. An improved system using GPS is also being investigated. Such a system provides enormous benefits by reducing considerably the need for hazardous and relatively expensive slow speed visual condition surveys as well as eliminating the inconsistency and subjectivity of manual interpretation. This development will be a major contribution to the effective monitoring and maintenance of flexible pavements. For the covering abstract, see IRRD E100539.

Language

  • English

Media Info

  • Features: References;
  • Pagination: p. 393-402
  • Serial:
    • Volume: 1

Subject/Index Terms

Filing Info

  • Accession Number: 00764285
  • Record Type: Publication
  • Source Agency: Transport Research Laboratory
  • ISBN: 82-519-1346-2
  • Files: ITRD
  • Created Date: May 28 1999 12:00AM