Research On Pavement Crack Recognition Method Based On Digital Image Processing

With the development of society and economy,the road network in China has been expanding,and the road transportation has been developing rapidly.However,the damage of road pavement influences the road service level and traffic safety.Therefore,traffic department pay more and more attention on the prevention and monitoring of pavement disease.Pavement cracks are the main initial performance in road surface diseases and are of great significance in the maintenance of highways.The traditional pavement crack detection system based on artificial vision has the defects of time consuming,high error and high risk of detection and implementation.It can not meet the needs of the current road development.With the rapid development of computer hardware,the automatic detection system of pavement crack based on image processing technology has become the focus of current research.Although this method can improve the efficiency of pavement crack detection greatly,there are still many problems,such as poor image denoising effect,poor adaptability of image enhancement algorithm and poor accuracy of image segmentation.Therefore,it is of great significance to study road surface crack identification method based on digital image processing.This paper takes the detection method of pavement cracks as the main line of research,and comprehensively analyzes the digital feature of the crack image by using a variety of digital image processing methods,in order to further develop and improve the existing pavement crack detection methods.First,aiming at the problem of image denoising,a median filtering method based on minimum standard deviation domain is improved.Secondly,in view of the problem of image enhancement,after exploring the correlation between the segmentation points and the average gray level of the image,an adaptive gray stretch segmentation method is proposed to achieve the effective stretch of the gray level of the fractured image,and combined with the improved high pass filter,the contrast degree of the pavement crack image is effectively improved.Then,in view of the problem of image segmentation,the advantages and disadvantages of edge detection and threshold segmentation are synthetically analyzed.A regional minimum threshold segmentation method combined with adaptive Canny edge detection is adopted.The edge of the adaptive Canny operator is superimposed on the segmentation result of the minimum threshold method of the region,so as to connect the pavement crack accurately.It is really connected.In addition,the pseudo crack and fracture fracture are found in the fractured surface of the fractured surface.In this paper,the pseudo crack region is eliminated according to the linear,length and area characteristics of the connected domain.By improving the traditional single directional structure elements,a multi direction structure element is constructed to perform the image execution morphology.The fractured region is connected by closed operation,and the types of the cracks are judged according to the number of connected regions and the projection information.By calculating the length and width of the linear cracks and the area and block degree of the network cracks,the characteristics of the image of the pavement cracks are extracted.Finally,it is proved that the algorithm has good practicability and feasibility for the recognition and extraction of the pavement cracks,combining with the analysis results of the image of the cracks in the actual section.To sum up,the automatic detection method of pavement crack based on image processing technology,which is based on image processing technology,has the characteristics of strong adaptability and good detection effect,and provides a reference for the deep research of pavement crack detection system.