Research On Model Of Crack Characteristic Parameters Calculation In Concrete Surface Image

In the construction of roads and traffics,concrete has been widely used because of its own active qualities.However,due to the influence of nature and the other factors,the surface of the concrete structure buildings are easily to form cracks.Using the crack analysis technology based on digital image processing to calculate the relevant characteristic parameters is helpful to the maintenance works,avoid the occurrence of safety accidents effectively.Therefore,the realization of automatic analysis the deseases of concrete structures surface has theoretical significanceand practical value.In this thesis,the core algorithm of crack analysis related technology has been deeply studied,the model of crack characteristic parameter calculation in concrete surface image has been established,the algorithm of image pixel singulation base on idea of template has been proposed,burr elimination algorithm has been improved,the crack classification algorithm has been optimized,the characteristic parameters calculation has been completed,and finally the crack analysis system of concrete surface image has been designed and implemented.The main researches of this thesis include:1.In this thesis,the thin algorithms of image have been deeply studied.Face to the problem of the non-single pixel width in the result of Zhang parallel fast thin algorithm,by analyzing the formation conditions of the protrusions,and combing with the template idea,four elimination templates were created to eliminate the protrusions,to make the results of thin algorithm single pixel in width.2.The burr elimination algorithms have been deeply studied in this thesis.For the buurs still exist after making the thin results single pixel in width,improved existing refinement algorithm,and followed the basic principle of retaining the longer branch as the main body information of the skeleton,each time choosed the shortest length branch to burr determination and elimination,ultimately achieved th single pixel in width and no burrs skeleton extraction.3.The crack classification algorithms have been deeply studied,based on the realization of crack skeleton extraction,optimized and improved the existing crack classification algorithm.On the basis idea of hierarchical classification,according to whether there have branch growth points to classify the cracks in first time.The second classification count the number of background connected area,and combin the idea of block division,divide the crack image in block,and then count the number of areas which contains the branch growth point.Finally,for different types of cracks choose the different geometric characteristic parameters have been calculated to indicate the degree of fracture damage.4.The effectiveness of the algorithms in this thesis have been verified by experimental analysis,and crack analysis system of concrete surface image has been designed and implemented.The system can process the input crack image in real time,and finally output the corresponding characteristic parameter calculation results,the system meets the requirements of high efficiency and high precision.Finally,the researchs and works abut this thesis have been summarized,and the follow research directions have beenpointed out,to preparing for further in-depth study in this thesis.