Design Of Pavement Crack Detection System Based On Image Processing

The repeated effects of the years of traffic,traffic loads,and environmental factors have made the problem of cracks on highway pavements increasingly prominent,it is the focus of current work designing a scientific,perfect and practical highway pavement crack detection system and improving the highway technical status assessment work level.It provides a powerful and convenient method for the study of pavement crack identification methods based on spatial distribution characteristics with the rapid development of computer vision technology and neural network theory.Therefore,the research can not only solve the current technical problems of the highway maintenance department,but also can be easily applied to the auto-driving system of automobiles which has great application value on pavement crack detection system based on computer vision technology.This paper first analyzes the technical indicators and overall structure of the existing road comprehensive inspection vehicles,and designs the software upgrade module structure for the actual inspection of the company.Secondly,it proposes a road crack identification method based on spatial distribution characteristics for road comprehensive detection.The vehicle system software is upgraded.This method uses the spatial distribution feature and the backpropagation neural network model to realize the automatic pavement crack classification.The direction coding(D-Coding)algorithm is used to encode the crack sub-section and extract the directional characteristics.At the same time,the Delaunay triangulation is adopted.The technology of the fracture area is analyzed by sub-technology and the density characteristics are extracted.Then the back-propagation neural network model is designed to classify the crack image of highway pavement.The computer simulation test results show that the classification accuracy of this method is over 92%,and the maximum recognition rate is92.9%.Finally,based on Visual Studio 2010 and Open CV3.4.1,this paper designs a software upgrade module for road surface crack detection system based on computer vision technology.The module realizes crack image enhancement,pavement crack space feature and density feature extraction,crack image recognition based on back propagation neural network,pavement crack image storage module,linear array camera calibration module,disease level and index calculation,etc.Related software flow diagrams and some key codes are finished.The test results of the pavement crack detection system based on digital image processing show that the software upgrading module of the road pavement crack detection system designed in this paper is qualified and achieved the expected goal.The system software can automatically detect the cracks in the highway pavement,and can calculate the relevant technical indexes.