Abstract
The development of automatic system to detect cracking information on pavement surface has been expected in order for the maintenance of a proper condition and repair planning of ever-expanding road-network. It is difficult to develop such a system that substitutes the conventional human-dependent interpretation and evaluation applying computer-based image analysing methodology due to a wided range of variation in images caused by the sundry pavement surface conditions and tough shooting environment, and such endeaver has not been successful in reaching the practical stage yet. This paper presents a cracking extraction system employing a combination of processing methodologies, which comprises of a process to enhance the linear continuity of cracking and a pavement surface density compensation process, as a processing algorithm for stable automatic extraction of pavement surface cracking.
The following results were obtained through the aforementioned set of processes.
1) The linear continuity of cracking could be enhanced and clearer cracking skeltal structure was obtained.
2) The fluctuation of the intensity of reflected light caused by uneven lighting and surface conditions could be compensated.
3) Moreover, a process to extract cracking lines stabely could be established.
