3D shape reconstruction of large specular surface

A novel three-dimensional (3D) reconstruction method based on fringe reflection technique for shape measurement of large specular surfaces is presented in this paper, which effectively integrates path integration technique with zonal wavefront reconstruction algorithm. The height information of specular surface obtained from cross-path integration can then be used as the initial value in a zonal wavefront reconstruction algorithm. This method not only has the advantages of global integration, but also enables user-friendly, high-speed operation. A specific iterative algorithm is adopted to improve the antinoise capability of the measuring system, which accelerates the rate of convergence significantly and even improves the accuracy of the reconstructed 3D surface. Moreover, the proper use of boundary contour extraction of the acquired images reduces the computational load of 3D reconstruction dramatically and hence achieves high reconstruction accuracy and enhances the surface integrity at the boundary. An ultraprecision, diamond-turned planar mirror with diameter of 150 mm has been employed to implement the system calibration. The reconstruction results of simulated and actual hyperbolic surfaces and the gauge blocks identify the validity of this new method. It is demonstrated that the measurement error is about 50 mu m with reconstruction points of 150 x 560 pixels of gauge blocks. (C) 2012 Optical Society of America