The Algorithm Of Cement Pavement Cracks Detection Based On Percolation Algorithm And Sample Variance

With the rapid development of the traffic infrastructure,China has become the largest country in the construction of highways,tunnels,bridges in the world,so the maintenance work has become gradually heavy.The crack is one of the most common diseases on cement pavement.If the crack is not repaired in time,the damage of the pavement will be further accelerated.With the rapid development of computer technology,Crack detection technology based on digital image processing can detect the crack quickly to overcome the limitations of traditional artificial crack detection technologies,and lay the foundation for the subsequent analysis of the crack information.Therefore,it is important for practical value for traffic maintenance work.This thesis studied the algorithm of crack detection based on percolation model deeply,and the key point of this thesis was to accelerate the time efficiency of crack detection.Firstly,the algorithm of cement pavement crack detection based on percolation algorithm and sample variance was proposed in order to improve the speed of percolation detection.Secondly,the crack similarity judgment algorithm was proposed to repair the fractured crack in the first percolation detection results.Thirdly,a cement pavement image crack detection system was designed and implemented to verify the feasibility and effectiveness of the proposed algorithm.The main research works include:1.This thesis analysed the current research situation of crack detection technology based on digital image processing.According to the technical difficulties of crack detection,the principle and algorithm of the key technology of crack detection were generalized.2.In this thesis,the algorithm of crack detection based on percolation model was researched deeply.Aiming at the problem that percolation detection had high precision but low efficiency,this thesis used crack region extraction algorithm based on sample variance to extract the crack regions,and proposed crack pixels extraction algorithm based on deviation processing to extract the crack pixels to reduce the number of pixels and enhance the timeliness of crack detection algorithm.3.The fractured cracks in the first percolation detection results could cause accuracy decrease.This thesis analysed the causes of crack fracture in percolation detection,and extracted the skeleton of the crack which based on Zhang fast parallel algorithm and eight direction chain code in concrete surface image.The similarity judgment algorithm of fracture zone was proposed,and the percolation processing was executed to repair the crack to avoid missed detection.4.This thesis designed and implemented a cement pavement image crack detection system which could process the cement pavement images in real time and output the crack binary image to achieve the requirement of crack detection rapidly and accurately.At the end of this thesis,the research works were summarized,and the future research directions and ideas were pointed out.