Zalevsky Zeev, Margalit Ofer, Vexberg Emanuel, Pearl Roy, Garcia Javier
School of Engineering, Bar-Ilan University, Ramat-Gan, Israel.
Appl Opt. 2008 Jul 1;47(19):D154-63. doi: 10.1364/ao.47.00d154.
In this paper we present two approaches for extracting the surface profile as well as obtaining 3D imaging of near field objects by usage of partial coherence and digital holography. In the first approach a light source with given temporal partial coherence is used to illuminate a near field object. The reflected light is interfered with the reference source. By computing the local contrast of the generated fringes one may estimate the 3D topography and the profile of the object. This approach extracts the 3D information from a single image, and its accuracy does not depend on triangulation angle like in fringe projection methods. The second approach is tomography based. There we illuminate the object from several slightly different angles, and for each we compute the wrapped phase using digital holography techniques. Combining the wrapped phase estimation from several points of projection allows calculating the unwrapped phase and therefore the true profile of even a phase-only object. Increasing the number of points of view decreases the relative error of the estimated profile.
在本文中，我们提出了两种利用部分相干和数字全息技术提取表面轮廓以及获取近场物体三维成像的方法。在第一种方法中，使用具有给定时间部分相干性的光源照射近场物体。反射光与参考光源发生干涉。通过计算所产生条纹的局部对比度，可以估计物体的三维形貌和轮廓。这种方法从单个图像中提取三维信息，其精度不像条纹投影方法那样依赖于三角测量角度。第二种方法基于层析成像。在这种方法中，我们从几个略有不同的角度照射物体，并且对于每个角度，使用数字全息技术计算包裹相位。结合来自多个投影点的包裹相位估计，可以计算出展开相位，从而得到即使是仅具有相位信息的物体的真实轮廓。增加视角的数量会降低估计轮廓的相对误差。
