多波长条纹投影三维测量中相位展开技术的研究

发布时间:2018-07-28 19:18
【摘要】:随着现代科学研究的深入和工业生产水平的不断提高,人们对于物体三维形貌准确测量的需求日益强烈,基于数字条纹投影的三维形貌测量技术作为新一代非接触式三维形貌测量技术,由于其高分辨率、无破坏、数据获取速度快、实现简单等优点而被认为是最有前途的三维形貌测量方法之一,已经被广泛应用于工业自动检测、产品质量监控、机器视觉、逆向设计、生物医学等众多领域。基于数字条纹投影的三维形貌测量技术是利用投影机向待测物体投影一组光强呈正弦规律变化的条纹,并用相机同时拍摄经物体表面形貌调制的变形条纹,然后利用条纹分析技术和相位展开技术从变形条纹中求解出包含物体表面三维信息的绝对相位,结合标定参数计算出待测物体表面的三维点云数据。其中,相位展开技术直接决定重构的三维形貌数据的准确性,是三维测量过程中关键的步骤之一。本论文详细介绍了基于数字条纹投影的三维形貌测量的基本原理、经典的条纹分析技术和相位展开技术,分析了现有的相位展开方法中存在的问题,然后针对传统的时间相位展开算法需要投影和拍摄的图片数目过多,实时性差等问题对如何在最小化所需图片数目的同时提升相位展开的抗噪性行了深入研究。首先通过对现有的多频率投影条纹相位展开算法的分析,特别是针对基于空间频率选择的双频率条纹投影相位展开算法中对条纹设计约束较大、相位误差容限不够高等问题提出一种基于空间波长选择的双波长条纹投影算法。论文详细介绍了该方法的基本原理和实现步骤,表明该方法无需复杂计算,仅是通过不等式的求解建立表征包裹相位图和条纹阶数之间一一对应关系的查找表即可快速完成相位展开操作;论文分析了所选两个条纹空间波长的约束条件和相位展开的相位误差容忍范围,指出所选两个条纹空间波长的最大公约数越大,则可容忍的最大相位误差就越大,并通过实验验证了该方法的有效性。双波长条纹投影相位展开算法减少了对投影条纹设计的约束、降低了相位展开过程的计算复杂度、提高了现有获取展开相位算法的速度、增强了抗噪性,是一种有价值的时间相位展开算法。通过对基于空间波长选择的双波长条纹投影相位展开算法的进一步研究发现,如果想要更大程度的提升算法的抗噪性,双波长法不得不增大投影条纹的空间波长,而空间波长的增加将会导致三维测量性噪比的降低,因此,文章提出了基于空间波长选择的三波长条纹投影相位展开算法,相对于双波长法能获得更好的抗噪性和测量性噪比。因此,当双波长算法能够从两幅六步相移技术得到的包裹相位图中成功恢复绝对相位(即仅需12幅条纹图)时,三波长算法可以从三幅三步相移技术得到的包裹相位图中成功恢复绝对相位,仅仅需要投影和采集9幅图片即可实现正确的相位展开,三维形貌测量的效率得到提高。文章同样对该算法的基本原理、空间波长选择的约束条件和抗噪性做了详细的分析论证。在基于空间波长选择的三波长条纹投影相位展开算法中,通过实验发现尽管该算法能够成功的恢复出绝对相位,但结果中依然存在少数规律性的错误恢复点,研究发现这些错误点的产生与空间波长的选择有关,归因于外差现象的出现,可以通过更合理的选择条纹空间波长来规避这些错误,但却会加大对条纹设计的约束。基于此,通过对相对误差率的计算我们提出了利用数据插值的方法修正相位展开的错误点或利用质量模板的方法移除这些错误点,实验证明这些处理规律性相位展开错误点的方法对含有庞大点云数据的三维测量结果影响很小。以上研究对基于空间波长选择的多波长投影条纹时间相位展开方法进行了完整的论证说明,为该方法在三维形貌测量中的应用奠定了理论和实验基础。在本文的最后对全文的研究工作进行了总结,并指出了后续研究的方向。
[Abstract]:With the deepening of modern scientific research and the continuous improvement of the level of industrial production, the demand for accurate measurement of three-dimensional shape of objects is becoming more and more intense. The 3D topography measurement technology based on digital fringe projection is a new generation of non-contact 3D topography measurement technology. Because of its high resolution, no destruction and fast data acquisition, the data acquisition speed is fast. It is considered to be one of the most promising methods of 3D topography measurement, which has been widely used in many fields, such as industrial automatic detection, product quality monitoring, machine vision, reverse design, biomedicine and so on. The 3D topography measurement technology based on digital fringe projection is a group of light intensity projection from projector to the object to be measured. The stripe, which is changed by the sine law, is photographed by the camera at the same time. Then the absolute phase of the three-dimensional information containing the surface of the object is solved by the fringe analysis technique and the phase unwrapping technique, and the three dimensional point cloud data of the surface of the object are calculated with the calibration parameters. It is one of the key steps in the process of three-dimensional measurement to directly determine the accuracy of the reconstructed 3D topography data. In this paper, the basic principle of 3D topography measurement based on the digital fringe projection, the classic fringe analysis technique and the phase unwrapping technique are introduced in detail, and the existing problems in the existing phase unwrapping method are analyzed. In view of the traditional time phase unwrapping algorithm, the number of pictures which need to be projected and photographed is too much, and the problem of real time is poor. This paper makes an in-depth study on how to improve the noise resistance of phase unwrapping at the same time to minimize the number of pictures needed. In the dual frequency fringe projection phase unwrapping algorithm for spatial frequency selection, a two wavelength fringe projection algorithm based on spatial wavelength selection is proposed. The basic principle and implementation steps of the method are introduced in detail. The phase unwrapping operation can be quickly completed by establishing a lookup table representing a one-to-one correspondence between the phase diagram and the order of the stripe. The paper analyzes the constraints of the wavelength of the two stripes and the range of phase error tolerance of the phase unwrapping, and points out the maximum common divisor of the two spatial wavelengths of the selected stripe. The larger the maximum phase error tolerable, the greater the maximum phase error is, and the effectiveness of the method is verified by experiments. The dual wavelength fringe projection phase unwrapping algorithm reduces the constraint on the design of the projection stripe, reduces the computational complexity of the phase unwrapping process, improves the speed of the existing phase algorithm and enhances the noise resistance. It is a kind of noise resistance. A valuable time phase unwrapping algorithm is obtained. By further study of the dual wavelength fringe projection phase unwrapping algorithm based on the spatial wavelength selection, it is found that if the enhancement of the noise resistance of the algorithm is greater, the dual wavelength method has to increase the space wavelength of the projection stripe, and the increase of the space wavelength will result in the three-dimensional measurement noise. As a result, the three wavelength fringe projection phase expansion algorithm based on space wavelength selection is proposed, which can obtain better noise and measurement noise ratio compared with the dual wavelength method. Therefore, when the dual wavelength algorithm can successfully recover the absolute phase (that is, only 12 stripes are needed in the package phase bitmap obtained from two six step phase shift techniques. " As a result, the three wavelength algorithm can successfully restore the absolute phase in the package phase graph obtained from the three three step phase shift technique. Only the projection and acquisition of 9 pictures can achieve the correct phase unwrapping, and the efficiency of the 3D topography measurement is improved. The basic principle of the algorithm, the constraint conditions of the space wavelength selection and the resistance are also given. In the three wavelength fringe projection phase unwrapping algorithm based on space wavelength selection, it is found that although the algorithm can successfully restore the absolute phase, there are still a few regular error recovery points in the result. The study found that the error points are produced and the space wavelengths are selected. Due to the appearance of heterodyne, we can avoid these errors by more reasonable selection of the fringe space wavelengths, but it will increase the constraints on the design of the stripes. Based on this, we propose a method to correct the error points of the phase opening by the method of data interpolation or remove this method by calculating the relative error rate. Some error points have been proved by experiments. The experimental results show that the method of handling the error points of the regular phase is very small on the three-dimensional measurement results containing large point cloud data. The above study is a complete demonstration of the multi wavelength projection fringe time phase unwrapping method based on the spatial wavelength selection. Based on the theoretical and experimental foundation, the paper summarizes the research work and points out the direction for further research.
【学位授予单位】:华中科技大学
【学位级别】:博士
【学位授予年份】:2016
【分类号】:TP391.41

【相似文献】

相关期刊论文 前10条

1 杨家骅;姜明启;;应用计算机分析光学条纹图的研究[J];交通与计算机;1992年05期

2 张海波,伍小平;二维相移逻辑云纹法──条纹图全自动处理的一种方法[J];实验力学;1994年03期

3 任传波,蒲琪,云大真,于万明;用图像运算技术对光弹性条纹进行倍增[J];光学学报;1999年05期

4 戴福隆,王朝阳;条纹图象的数字化自动分析处理技术之一:条纹中心法[J];光子学报;1999年08期

5 张辉;王东鹤;穆治亚;;十六步相移技术在弹痕条纹图分析中的应用[J];光电工程;2013年01期

6 李庆丰,丁淑杰;图像变异条纹族间距的一种自动确定方法[J];黑龙江商学院学报(自然科学版);1998年02期

7 马广云,申功p,

本文编号:2151321


资料下载
论文发表

本文链接:https://www.wllwen.com/shoufeilunwen/xxkjbs/2151321.html


Copyright(c)文论论文网All Rights Reserved | 网站地图

版权申明:资料由用户c303d***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱[email protected]