摘要
在基于相位信息的三维面形测量中,为了得到正确的相位展开而引入的截断线和为了便于三维拼接而附加在物体表面的标志点都会造成其覆盖区域的相位缺失,需要进行人为地插补修复。一种新的缺失相位插补方法——八方向梯度估算法,通过计算待插补点周围八方向梯度和其八邻域内的相位平均值,可估算出缺失的相位。与其它常用缺失相位插补方法相比,该方法运算速度快、插补效果好,运算过程无需人为参与,便于计算机自动实现,尤其适合于插补倾斜相位面上的较大标志点,有助于提高相位重建质量和在此基础上的三维面形重建质量。
In optical three-dimensional (3D) measurement with the structured light, the branch cut and marked point are usually used to avoid the error in phase unwrapping and to carry out 3D registration respectively. Covered by the branch cut and marked point, the absent local data in the unwrapping phase need factitiously interpolate. A novel interpolation algorithm gradient estimate algorithm, which recovers the missing phase with the average of the nearest neighbors and the gradient in eight directions. Compared with other interpolation algorithms, this new algorithm works fast and gains well interpolated result especially when it deals with marked point on inclined phase surface. It needn't any predefined parameters in the interpolation process. So the gradient estimate algorithm can he widely used in missing phase interpolation for optical 3D measurement application in many fields.
出处
《光学技术》
CAS
CSCD
北大核心
2009年第4期502-505,509,共5页
Optical Technique
关键词
光学测量
插值
相位展开
截断线
三维拼接
标志点
optical measurmaent
interpolation
phase unwrapping
branch cut
three-dimensional registration
marked point