The High Reflective Surface 3D Reconstruction Algorithm Researching Based On The Fringe Reflection Technique

With the high-speed development of modern manufacturing,the specular and mirror-like surface that own the high reflective surface features have been more and more apply to automobile manufacturing,aerospace science and technology,national defense industry,medicine,science and so on,through its own various superior features.The measurement studying of reflecting surface is increasingly important.The3 D measurement method that based on the fringe reflection technique is a new measuring technology which is based on the computer vision and fuse the information optics.It has become the common method of the high reflective surface measuring technology through the advantage of non-contact,good real-time and high precision.We combine the cross path integration with iterative compensation to reconstruct the object surface shape based on the fringe reflection,achieving the more stability accuracy and improving the reconstruction algorithm convergence speed.In this paper,the improved algorithm use cross path integral method to obtain the initial surface relatively rough shape at first,and regard it as the iterative initial surface shape of the iterative compensation algorithm in the improved algorithm's second step,finally use the iterative to reconstruct the 3D shape of the high reflecting surface and improve the measurement accuracy.In this paper,we study the content mainly includes:1.Proposing the improving algorithm that combine the cross path integration with iterative compensation,setting up the high reflective surface measurement system based on fringe reflection technique and briefly describes the measuring principle,discussing various 3D surface reconstruction algorithms that based on the gradient data,analyzing the advantages and disadvantages of various reconstruction algorithms and introducing the specific process of the improved algorithm.In order to eliminate the influence of system error and other external factors on the experimental results,calibrating the CCD camera of the measure system and the distance from side to screen.2.Using MATLAB to simulate the surface reconstructing of the hyperboloid and peaks function,comparatively analysising the cross path integral algorithm,the iterativecompensation algorithm and the improved algorithm.Actually measuring the concave spherical mirror that the curvature radius is 1m and size is 80mm×80mm.Simulation and experimental results prove the feasibility and superiority of the improved method.