Research On Key Technologies Of Pavement Crack Detection Based On Structured Light

With the rapid development of highway in our country,the pavement crack detection has become more and more important in our country highway construction.In view of the issue that the pavement crack detection system based on two-dimensional image cannot overcome the influence of shadow,oil,light conditions and other factors on pavement crack detection,according to the structure light detection technology has high accuracy,the strong ability to overcome light environment and obtain 3D information,this paper studies the key technologies of the pavement crack detection based on structure light.Firstly,this paper studies the pavement crack detection system based on structured light by the demand of the pavement crack detection system.Aiming at the problem of measurement error caused by bump of the system,measurement error models were established under different turbulences with analyzing the influence of the attitude on measurement error and the optimal structure of the sensor system was determined by the error minimization principle.The inspection model was established in order to solve the problem that the turbulence influences on the depth resolution and the design method of sensor structure parameter was put forward for technical requirements of the system,then the structure parameters of the sensor were optimized by the error analysis theory.The experimental results show that the sensor structure parameters design method can optimize the sensor structure parameters and improve the accuracy of the system.Secondly,in view of the issue that the structure light plane calibration method needs complex process and high cost,according to the local linear characteristics of the line structure light sensor imaging,a direct calibration method was proposed.The experimental results show that the calibration method proposed in this paper is simple and accurate.Thirdly,in view of the issue that the line structured light center extraction algorithm cannot satisfy the accuracy and real-time performance simultaneously,according to the two-dimensional gray centroid can overcome the influence of stripe direction,the two-dimensional gray gravity method with adaptive window put forward,and a classification method was proposed for solving the actual engineering light off line problem.The experimental results show that the error of this method is reduced by 36% compared with the traditional gravity method,and the time consuming is only about the improvement of the method by 38%.Finally,in view of the issue that the contrast of pavement depth image is little and pavement has different morphology interferences,according to the characteristics of the crack of depth image and the characteristics of the crack edge of 2D pavement surface image,a preprocessing algorithm based on piecewise function was proposed for the crack image.Then this paper studied the multi scale rough extraction algorithm based on Hessian matrix.To solve the problem that the rough extraction algorithm lead to the local discontinuous cracks and enhances the noise,according to the continuity direction of the local crack and the discreteness of the noise,this paper studied the accurate extraction algorithm based on the constraint of linear structure.The experimental results show that the proposed method is suitable for the extraction of pavement cracks with different scales and degrees,compared with the method of document [57],the extraction integrity is improved by more than 10%,and the extraction correct rate is increased by more than 2%,compared with the method of document [69],the extraction integrity is improved by more than 6%,and the extraction correct rate is increased by more than 1%.