A man-machine balanced rapid object model for automation of pavement crack sealing and maintenance

A number of studies during the last few years have discussed automated crack detection and mapping using digital image processing technologies in roadway maintenance and rehabilitation. Many recent studies have applied dig- ital image processing to the recognition or sealing of cracks in pavement. There have been great discrepancies, how- ever, among various segmentation methods that extract crack types and locations or classify the extent of cracking. Since all sensing systems also produce some spurious data and experience noise due to the varied topological and color conditions of the pavement surface, accurately mapping and representing the pavement cracks to be sealed using such segmentation methods would be even harder. This paper illustrates an innovative machine vision algorithm developed for accurate crack mapping and representation in the University of Texas (UT) automated road maintenance machine (ARMM). The paper mainly focuses on illustrating the detailed logic and descriptions of the algorithm. Efficiency eval- uation results of the ARMM man-machine balanced crack mapping and representation process, including the line- snapping and path-planning functions, are also shown. Using the algorithms as an edge-describing tool can have broader applications in automation of infrastructure maintenance and inspection of civil works and in the domain of digital image processing.