A New Technique for Automatic Detection and Parameters Estimation of Pavement Crack

Pavement condition evaluation is a significant part of a good pavement management system for effective maintenance, rehabilitation, and reconstruction decision-making. One of the key components of pavement condition evaluation is the quantification of pavement distresses data. Cracking is the main form of early pavement distresses. Cracking of pavement affects road condition, driving comfort, traffic safety, and consequently reduce pavement service life. Once initiated, cracking increases in extent and severity and accordingly accelerates the rate of pavement deterioration. Therefore, the awareness about crack type, extent, and severity is essential to evaluate pavement condition and to determine timing and cost of pavement maintenance. Digital image-based automated pavement evaluation has been gradually replacing the manual pavement evaluation due to its improved efficiency and safely operating. In this paper, we are presenting a novel reliable automated pavement assessment system based on image processing techniques and machine learning methods. The proposed system has the ability to i) identify crack, ii) extract crack parameters, and iii) report the type, extent, and severity level of that crack in an output file. Actual pavement images were used to verify the performance of the proposed system. The results clearly demonstrated that the proposed system was able to automatically and effectively identify crack type and efficiently extract crack parameters from pavement images. Such information can be used by public road agencies to define maintenance plans and assist in pavement management decision-making, in accordance with real pavement condition.