摘要
针对基于结构照明的三维面形测量中的位相展开问题，提出了基于复Morlet小波分析的直接从条纹中提取自然相位的算法，并对其进行了理论分析，计算机模拟和实验，研究证明了可行性。该方法从一幅条纹图像中不需要进行相位展开就能够得到准确的相位分布方法，并且避免了传统相位方法所必须的复杂相位展开过程．当待测物体变化率比较大的时候，用传统傅里叶方法恢复得到的误差是-3～3rad，而采用本文方法得到的误差是-2．5-0．5rad，且对于顶部平坦部分基本没什么误差。对比分析表明，本文方法的精度要高于传统傅里叶方法。
Almost all of the three-dimensional measurement methods based on the structured light illumination are faced with the phase-unwrapping problem. A new direct phase-extracting algorithm based on complex Morlet wavelet analysis is discussed. The theoretical analysis and experiments prove that the tested phase distribution can be accurately obtained from a single fringe pattern without phase-unwrapping step by this method. In addition, comparing our method with the traditional FTP method, it is found that the measurement result of the new method is more accurate when a tested object with a great gradient is detected. As is shown in the computer simulations, the measurement error is decreased to [-2.5rad, 0.5rad] by our method, while the error is [-3rad, 3rad] by the traditional Fourier transform method.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2007年第4期73-76,119,共5页
Opto-Electronic Engineering
基金
国家自然科学基金资助项目(60527001
60677028)
关键词
小波分析
三维面形测量
相位分析
计算机模拟
Wavelet analysis
3-D shape measurement
Phase analysis
Computer simulation
