Research On Curved Surface Measurement Method Based On Line Structured Light

With the rapid development of China's manufacturing industry,products with complex curved surfaces have been widely used.The functioning of these products greatly depend on the manufacturing accuracy of the curved surface.Therefore,it is necessary to measure the curved surface and evaluate the manufacturing accuracy of the curved surface with high accuracy.At present,the measurement of curved surface is mainly carried out by human or CMM(Coordinate measuring machine)with relatively low measurement efficiency.Moreover,these measurement methods have certain limitations depending on the curved surface shape.Therefore,this paper designs and implements a non-contact surface measurement system based on the three-dimensional measurement technology of line structured light and studies the related key technical steps.Firstly,to address the problem that the measurement of the common monocular structured light sensor could be incomplete when measuring the surface with large contour fluctuation due to the obstruction of light,this paper adopts the binocular structured light sensor.The sensor uses two cameras to collect the linear structured light stripe image at the same time under different viewing angles to reduce the influence of occlusion.Secondly,this paper describes the components of the hardware experiment platform and measurement process of the surface measurement system.The integration of camera calibration and line structured light plane calibration is realized based on Zhang's calibration method,which lays a foundation for the subsequent acquisition of 3D data.In order to solve the problem of large deformation of line structured light stripe and difficulty in the extraction of the center caused by irregular surface contour,the improved gray center of gravity method is adopted.The experimental results are compared with the traditional extraction method.The results show that the improved gray center of gravity method has strong robustness for the shape change of light stripe,and the extraction efficiency can also meet themeasurement requirements.Thirdly,ICP algorithm based on k-d tree acceleration is applied for the fast registration and fusion of two groups of point clouds from the perspectives of two cameras.To solve the problem of redundant points in the fusion point cloud,an improved bounding box algorithm is used to simplify the point cloud quickly.Based on this,the 3D model of the measured surface is reconstructed using Poisson surface reconstruction algorithm.Finally,based on the platform of Visual Studio 2015 and QT,Open CV and PCL open source libraries,the software of the measurement system is developed.Through the measurement on a gypsum model with complex surface,the feasibility and effectiveness of the measurement system are verified.