Pavement Distress Detection Method And Equipment Technology Based On Image Feature

With the progress of science and technology development,the construction of expressway is getting more and more attention,road repair work has become the focus of attention,how to quickly and accurately detect defects in pavement crack problems,has become a priority among priorities of the problem we need to solve the.At present,the level of road detection system used in our country is still very backward,low efficiency,great danger,and obviously behind other countries.However,in recent years,due to the popularity and development of artificial intelligence technology,both at home and abroad have begun to research and Practice on the pavement crack detection project,and strive to develop a system for non-contact measurement of pavement,this system is mainly for a large-scale detection system of expressway.It is the most suitable way to detect the road in the safety factor and the real time compared to the original technology.But during the detection process,on the way to the pit,oil,shadows,signs and other uncontrollable factors,affecting the precision and efficiency of detection,interference and light,uneven material particles between random noise and other factors,which makes the system faces many technical problems.This paper is mainly for the non contact detection system and algorithm of expressway pavement to carry out research,comprehensive domestic and abroad on the crack detection theories,put forward a kind of convenient operation and can realize rapid detection of road detection algorithm,it can solve the noise interference problem in the detection of road surface crack.This paper mainly completes the work:Firstly,the image acquisition system of vehicle mounted road detection preprocessing for uneven illumination,not obviously damaged road signs and other factors,road,the theoretical analysis and mathematical calculation,finally compare.Finally we proposed an adaptive algorithm to analyze a series of problems,can restrain the noise area effectively,keep the target information of the whole area of the background and target information,complete segmentation.Secondly,to complete the pavement crack image preprocessing image edge detection,need to achieve a short period of time to extract the requirements of the target information area.The extraction method of non negative region of the target information,the improvement in the integrity of the target information extraction,and greatly accelerate the speed of running the program,improve the efficiency,also proposed adding method connected component labeling denoising,can traverse the entire image,and to shorten the running time,and can make the removal the noise of the maximum crack information can be extracted effectively.