Font Size: a A A

Design And Realization Of Vehicle-mounted Pavement Crack Identification System Based On Image Processing

Posted on:2014-03-30Degree:MasterType:Thesis
Country:ChinaCandidate:J LiFull Text:PDF
GTID:2268330422964693Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
With the rapid develpoment of highway construction, highway Maintenance hasbecome a more and more important aspect. How efficient, convenient and fast to detectthe cracks on the highway and to develop appropriate strategies for road maintenance forroad maintenance personnel is becoming a difficult problem they faced by. In the past,road maintenance methods are in accordance with either manual or semi-automatic way tocomplete the pavement crack detection and classification. This approach takes a lot ofmanpower and material resources, but also vulnerable to man-made factors. Therefore,research and design on automatic pavement crack identification system has importanttheoretical and practical significance.This paper reviewed a number of literatures on the road crack classification methods, andclassifies them into two major categories: unrepaired crack and repaired crack. Theunrepaired crack is further divided into transverse cracks repair cracks, longitudinal cracks,block cracks and alligator cracks. This paper carried out in-depth research for the problemof the repaired crack detection and the road crack classification.For the problem of repaired crack detection,we use the variance and graycharacteristic of the repaired crack. Through the image preprocessing,image segmentationand repaired crack extraction steps,we will obtain the crack Tagged Image. For theproblem of unrepaired crack classification, we use the differences of them on the crackdirection and the distribution for crack classification. Using2D feature mapping,Delaunay triangular subdivision and other methods, we will calculate some relevantcharacteristic parameters of these features to characterize differences of the crack toachieve the purpose of classification. Finally, the use of Matlab/GUI demo platformsimulation tool design algorithms and algorithm simulation experiments, and byevaluating the experimental results verify that the algorithm is efficient.The algorithm proposed in this paper is designed to uphold the principle of practicalapplication and reflects the suitability of certain algorithms.
Keywords/Search Tags:Image pre-processing, Feature extraction, Detection of pavement crack, Fracture types
PDF Full Text Request
Related items