Rapid Identification Technology Of Pavement Crack Based On Image Processing

In recent years,the development of highway traffic system in our country is very fast;highway mileage has increased year by year;city road system arrangement also gradually towards humanized and scientific;traffic conditions towards convenient,high-efficiency and safety,fulfilling the travelers' traffic demand.However,with the efficient and convenient traffic conditions,the road will have to bear the traffic volume increasing,especially for high-speed and heavy load vehicles,various pavement diseases emerges,the most common type of diseases is crack.Therefore,the highway system managers and engineers are paying more and more attentions to pavement inspection,repair and maintenance.Asphalt pavement crack detection is one of the key factors in pavement management system.Usually,in artificial visual inspection is low efficiency and dangerous,and the detected data information is discrete.Therefore,in order to achieve efficient and fast pavement distress detection,automatic detection system emerges as the times require.This paper mainly engages in research of rapid identification of pavement crack,discussing the automatic detection technology of pavement crack based on image processing,mainly proposing the method of pavement crack image preprocessing,image segmentation and feature parameter calculation.This article mainly completed several aspects of work as follows:According to the usage of domestic and foreign pavement crack automatic detection equipment,infroducing the general structure design principle and hardware constitution of line laser pavement crack detection system.This paper based on the traditional spatial domain image smoothing methods,proposing the smoothing method based on multi-directional linear region,studying the classical image segmentation algorithm,and putting forward the improvement measures of classical image segmentation method,laying down the partially adaptive threshold segmentation method in complex images.The author also verifies the superiority of the image processing method in this paper through examples.The results reveal that the method in this paper can output satisfactory segmentation images.Finally,around the binary images,the author carries out the research of crack types identification method,and feature extraction method,such as area,length and width.Through analysis of typical cases,the author introduces the working process of the software developed in this paper;by software test results of 100 road images,the author verify the feasibility of the proposed method in this paper and the stability of the software system developed in this paper.