Welcome to the National Transport Library Catalogue

Normal view MARC view

Pilot for automated detection and classification of road surface degradation features. Final report Javidi, Bahram et al

By: Publication details: Storrs, CT University of Connecticut, [2003?]; Connecticut Transportation Institute, ; Connecticut Department of Transportation, Description: 31 sSubject(s): Bibl.nr: VTI 2005.0885Location: Abstract: This report deals with the detection and classification of pavement cracks. Currently, ConnDOT is using Wisecrax® which is a commercial product supplied by Roadware. However, Wisecrax® has some shortcomings. We proposed seemingly different techniques based on wavelet and Hough transforms; since the problem at hand must be analyzed in both the spatial and frequency domains. Two main algorithms had been developed for elimination of false detection, and for more accurate detection. The proposed techniques require less user intervention, and showed promising results. Some road images had been analyzed, specifically those which Wisecrax® did not analyze well. Since we tested our algorithms against highly compressed road images, some parts of the cracks were not detected, causing errors in crack length measurements.
Item type: Reports, conferences, monographs
Holdings
Current library Call number Status Date due Barcode
Statens väg- och transportforskningsinstitut Available

This report deals with the detection and classification of pavement cracks. Currently, ConnDOT is using Wisecrax® which is a commercial product supplied by Roadware. However, Wisecrax® has some shortcomings. We proposed seemingly different techniques based on wavelet and Hough transforms; since the problem at hand must be analyzed in both the spatial and frequency domains. Two main algorithms had been developed for elimination of false detection, and for more accurate detection. The proposed techniques require less user intervention, and showed promising results. Some road images had been analyzed, specifically those which Wisecrax® did not analyze well. Since we tested our algorithms against highly compressed road images, some parts of the cracks were not detected, causing errors in crack length measurements.