Three Dimensional Dectection Technology Of Tunnel Inner Wall

In order to alleviate the shortage of urban land resources and improve the urban environ-ment,my country has paid more and more attention to the development of underground space,and the development capacity has become more and more mature.Among them,tunnels ac-count for a large proportion of underground space development.According to statistics,by the end of 2020,the total length of railway tunnels in operation in my country will be about 19,630kilometers.The increasing scale of the tunnel is accompanied by the emergence of safety prob-lems.Due to the high cost and low efficiency of the existing automated testing equipment,it cannot meet the requirements of actual use.Therefore,the security detection of the tunnel is still mainly based on manual detection.However,manual detection is prone to errors caused by factors such as short tunnel skylights,insufficient lighting,and manual negligence.The 3D scanning technology based on structured light is a hot research topic in 3D in-spection at present.Among them,the 3D scanning technology based on monocular structured light has the advantages of high precision and fast speed,and has attracted the attention of domestic and foreign researchers,but point cloud registration needs to be performed on the point cloud data after scanning.Perform point cloud registration on point cloud data.Based on the analysis of the existing 3D scanning model,this paper proposes a 3D scanning detection model based on annular structured light to scan and detect the inner wall of the tunnel.The closed-loop characteristics of annular structured light are cleverly used,and no point cloud is required after scanning.With registration,the internal point cloud data can be obtained in a single scan,and the 3D point cloud data can be used to conduct safety inspection of the tunnel.The specific work content is as follows:（1）The establishment of a three-dimensional scanning detection model for a ring-shaped structured light sensor.The advantages and disadvantages of the current common structured light visual scanning models are analyzed and summarized,and a 3D scanning detection model for the inner wall of the tunnel based on annular structured light is proposed in combination with the characteristics of the tunnel itself.The scanning equipment of this model is composed of cameras,annular structured light and track trolleys?a three-dimensional detection math-ematical model is established according to the characteristics of the tunnel itself,and it is verified by experiments.It realizes convenient,simple,dynamic,efficient,and high-precision3D scanning inspection.（2）Camera parameter calibration and structured light fringe center extraction.The camera parameter calibration adopts Zhang Zhengyou’s camera calibration algorithm.The commonly used structured light fringe center extraction algorithms are compared experimentally.Finally,according to the characteristics of the annular structured light fringes and the experimental results,the Hessian matrix method is selected to extract the light fringe center of the annular structured light fringes.（3）Calibration of the scanning model of an annular structured light sensor.The calibration of this part is mainly divided into two parts:the checkerboard plane equation parameters(6,(7,(8,(9 calibration and the annular structured light cone equation parameters(6₁,(6₂,(6₃,(6₄,(6₅,(6₆,(6₇,(6₈(6₉calibration.The annular structured light is projected onto the checkerboard plane with the calibration target,and the three-dimensional coordinates of the checkerboard corner points in the camera coordinate system are calculated,and the calibration target plane is calibrated using these corner coordinates.The three-dimensional coordinates of the annular structured light in the camera coordinate system are calculated by using the method of line-surface intersection,and the parameters of the annular structured light are calibrated using the calculated three-dimensional coordinates of the annular structured light.After the experimental system was built and calibrated,the physical models of differ-ent materials were scanned and found that the absolute error between the scanned point cloud data of the plastic flowerpot and the actual data was 2.045mm,and the relative error was0.314%?the absolute error between the scanned point cloud data of the carton and the actual data is 2.804mm,and the relative error is 0.596%?the absolute error between the scanning point cloud data of the corridor and the actual data is 17.892mm,and the relative error is 0.568%.The acquisition rate of point cloud reaches 20000p/s.Experiments show that the method has the characteristics of high precision,high efficiency and simplicity,and good robustness.The three-dimensional point cloud data obtained by scanning are tested and verified by experi-ments,which can effectively detect the protrusions on the inner wall,and the method is simple and effective.The method can be used in the three-dimensional measurement and inspection of the inner wall of the pipeline or tunnel.