一种时域平面扫描三维成像算法研究

提出了一种基于平面扫描模式下的三维成像时域算法。该算法在平面网格点上获取扫频幅相信息,经过相位补偿后变换到球面上,然后利用傅里叶逆变换将频域数据沿投影线转换为时域数据,对应球面的不同空间角度在投影线上作线性插值,最后通过二维角域积分获得目标的三维图像。仿真结果表明该算法聚焦效果良好。在微波暗室内构建了一套平面扫描三维成像系统,采用该算法处理后能获得高分辨率三维微波图像,成像的空间位置误差优于1 cm。     

基于DLP数字微镜的结构光三维扫描系统

为了提高现代化生产中的效率和满足三维扫描的需求,设计了一种能够准确提取物件表面三维轮廓信息的扫描系统。系统采用结构光三维成像方法,先通过普通白光将光栅投影到被测物体表面,利用工业相机采集变形光栅图,再根据变形光栅图像中的灰度值变化,用傅里叶变换轮廓法解调出三维坐标信息。实验结果表明,使用傅里叶变换轮廓法重构可获得效果理想的三维点云,其为三维轮廓扫描提供了实验方法,同时也为提高点云精度提供参考。     

鲤鱼尾鳍瞬时三维形态测量

提出了鱼在自由游动状态下尾鳍实时三维形貌的测量方法。将影栅云纹投射在鲤鱼尾鳍表面,产生包含三维形态信息的云纹,由高速摄像机采集瞬时序列变形条纹,运用傅立叶变换轮廓术对序列图像进行处理,并将相位解包裹方法用于动态过程瞬时三维相位场的相位展开中,重建了鲤鱼尾鳍自由游动的瞬时三维形态,再现了尾鳍的连续摆动过程。     

Study of the direct Fourier method for the three-dimensional reconstruction of objects in the case of a missing angular data range

Some practical and theoretical considerations on the three-dimensional reconstruction of an object from its two-dimensional projections by the direct Fourier method have been obtained using computer generated images. The procedure applied has been designed to study the specific problems derived from the collection of images in transmission electron microscopes (TEM), making special emphasis on images of biological objects. The study of the calculated point spread function (PSF) of the experimental system and the mean and variance profiles of the reconstructions have provided useful criteria to assess the quality and the interpretability of three-dimensional reconstructions of real data.     

A generalized method for analysis of three-dimensional extrusion of arbitrarily-shaped sections

Despite increasing demand for and application of three-dimensional extrusion of various sections through continuous dies, so far very little work has been done by systematic and general analysis to predict the plastic flow properly. For effective die design, efficient design method and the related method of theoretical analysis are required for extrusion of complicated sections. In generalized three-dimensional extrusion of sections through continuous dies, a new method of die surface representation, using blending function and Fourier series expansion, is proposed by which smooth transitions of die contour from the die entrance to the die exit are obtained. The flow patterns as well as the upper-bound extrusion pressures are obtained on the basis of the derived velocity field. The effects of area reduction, product shape complexity, die length and frictional condition are discussed in relation to extrusion pressure, the distorted grid pattern and distribution of the final effective strain on the cross-section of the extruded billet. As computational examples for arbitrarily-shaped products rounded rectangles and ellipses are chosen for the extruded sections. Experiments are carried out for aluminum alloys at room temperature for a rounded square section and an elliptic section. In order to visualize the plastic flow, the grid-marking technique is employed. The theoretical predictions both in extrusion load and deformed pattern are in good agreement with the experimental results.     

SPARK,a kernel of software programs for spatial reconstruction in electron microscopy

SPARK, an acronym for ‘SPAtial Reconstruction Kernel’, is the nucleus of a software library being developed for the three-dimensional (3-D) reconstruction of objects observed by the electron microscope. A unifying concept is used: the Fourier transform, known in several central sections, is resampled to obtain a 3-D Cartesian grid, which is inverted by a fast Fourier transform. This technique is used for both single-axis tilting (of 2-D periodic layers or of isolated objects) and for the random conical-tilt technique. The principles that make it possible to recover Cartesian grids in the two different geometries are illustrated and some preliminary results are reported. SPARK resamples the Cartesian grids with the use of a fast and efficient algorithm of Shannon interpolation developed by the authors. Compared to back-projection techniques the method shows a considerable improvement in execution time with no sacrifice in accuracy; it therefore allows the effects of a variety of parameters in a given reconstruction to be scrutinized in a reasonable time. Some new possibilities and future extensions of the library are briefly outlined.     

一种应用标准刻线与CCD直接获取运动位移的方法

将标准刻线应用到对微细物体进行高精度定位的交流伺服驱动控制系统中,提出了一种新的直接获取运动位移的方法。调整光路使CCD获取清晰的光栅图像,对图像进行分析处理,计算出栅距与像素之间的映射关系,再对工作台移动前后的光栅图像进行对比,采用基于李萨如图的相位差比率算法获取它们的像素位置差,从而计算出工作台的实际运行位移。试验结果表明,该方法能简单快速地实现微电子加工领域亚像素级的测量。     

基于LabVIEW的光纤傅里叶变换光谱仪数据处理技术

为了提高编程效率与增强光谱仪数据分析处理能力,结合LabVIEW软件在信号处理方面的优势,利用虚拟仪器技术实现了光纤傅里叶变换光谱仪干涉图处理与光谱复原。采用均匀抽样算法,利用由光纤Mach-Zehnder干涉仪输出的参考光干涉图过零点取样,实现对测试光干涉图的等光程间隔采样,消除了由压电陶瓷非线性光程调制引起的误差。采用求摸法复原出光谱图,自动消除了相位误差。利用LabVIEW软件完成了对宽带ASE光源的干涉图数据采集、显示与处理并最终得到光谱图。结果表明,与光栅光谱仪测得的光谱相比,光谱线型趋于一致。     

莫尔偏转法测量相位物体时其Talbot象面形状的确定

本文利用傅里叶光学原理,从理论上分析用莫尔偏转法测量相位物体时,第一块光栅形成的Talbot象面的形状分布及第二块光栅应放置的位置以及此时形成的莫尔条纹的光场分布情况。     

条纹投影三维形貌测量的变分模态分解相位提取

针对条纹投影三维形貌测量涉及的相位提取,提出了一种基于变分模态分解的单幅条纹投影相位提取方法。通过建立变分模态分解模型和极小化变分模态分解将单幅投影条纹图分解成背景部分、条纹部分和噪声部分。然后对得到条纹部分进行Hilbert变换和反正切变换得到包裹相位;对其进行质量导向相位解包裹和Zernike多项式去载频得到解包裹相位。将该方法与Fourier变换、连续小波变换进行了对比,结果显示:本文提出的相位提取方法相位误差为3.14×10-4,小于Fourier变换和连续小波变换方法对应的误差3.30×10-4和6.52×10-4。模拟和实验结果表明:本文提出的方法在处理具有边缘信息投影条纹图时具有优势,能够提取出更准确的相位信息,可有效地用于含边缘不连续和突起的三维物体测量。     

Linear phase imaging using differential interference contrast microscopy

We propose an extension to Nomarski differential interference contrast microscopy that enables isotropic linear phase imaging. The method combines phase shifting, two directions of shear and Fourier‐space integration using a modified spiral phase transform. We simulated the method using a phantom object with spatially varying amplitude and phase. Simulated results show good agreement between the final phase image and the object phase, and demonstrate resistance to imaging noise.     

一种基于无衍射光栅的三角测量的定标方法

基于光栅的三角测量方法是将光栅投射到被测物体表面,CCD接受由于物体表面高度起伏产生的变形光栅像,根据光栅变形量与高度的关系,恢复被测物体的三维形貌。提出了一种基于无衍射光栅的三角测量的定标方法。该方法通过空间坐标变换,方便地求出摄像机的内外参数,从而获得被测物体的三维信息。     

采用非规则标识点过程的LiDAR点云数据目标提取

针对LiDAR点云数据目标投影几何的非规则性,提出非规则标识点过程的LiDAR点云数据目标提取方法。首先,在投影平面上定义随机点过程,利用其随机点定位该平面上的目标投影,对每一随机点生成一组节点集以建模该目标投影几何,作为目标标识;假设地物目标高程值服从独立同一高斯分布,从而得到LiDAR点云数据高程测度模型;在贝叶斯理论架构下建立目标几何提取模型,并结合可逆跳变马尔可夫链蒙特卡罗(Reversible Jump Markov Chain Monte Carlo,RJMCMC)算法模拟后验分布以及估计各参数;最后根据最大后验概率准则,求解最优目标提取模型。采用提出方法对LiDAR点云数据进行目标提取,根据实验结果可以看出,算法得到的检测精度均达到80%以上,最高精度为99.43%,得到了较好的检测结果。本文将传统的规则标识点过程拓展到非规则标识点过程,可以有效拟合任意形状目标几何。定性和定量的实验结果表明了该方法的可行性、有效性和准确性。     

光栅在驱动控制系统运动位移测量中的应用

利用计量光栅的栅距作为基准刻线,对二维精密工作台的运动位移进行测量,提出了一种新的位移测量方法。通过CCD(charge coupled device)获取光栅刻线图像,对图像进行分析处理,计算出栅距与图像像素之间的映射系数,再通过对工作台移动前后光栅图像的对比,采用基于李萨如图的相位差比率算法获取它们的像素位置差,从而计算出工作台的实际运行位移量。     

Reconstruction of optical pathlength distributions from images obtained by a wide-field differential interference contrast microscope

An image processing algorithm is presented to reconstruct optical pathlength distributions from images of nonabsorbing weak phase objects, obtained by a differential interference contrast (DIC) microscope, equipped with a charge-coupled device camera. The method is demonstrated on DIC images of transparent latex spheres and unstained bovine spermatozoa. The images were obtained with a wide-field DIC microscope, using monochromatic light. After image acquisition, the measured intensities were converted to pathlength differences. Filtering in the Fourier domain was applied to correct for the typical shadow-cast effect of DIC images. The filter was constructed using the lateral shift introduced in the microscope, and parameters describing the spectral distribution of the signal-to-noise ratio. By varying these parameters and looking at the resulting images, an appropriate setting for the filter parameters was found. In the reconstructed image each grey value represents the optical pathlength at that particular location, enabling quantitative analysis of object parameters using standard image processing techniques. The advantage of using interferometric techniques is that measurements can be done on transparent objects, without staining, enabling observations on living cells. Quantitative use of images obtained by a wide-field DIC microscope becomes possible with this technique, using relatively simple means.     

Digital double-pulse holographic interferometry using Fresnel and image plane holograms

Different arrangements for digital double-pulse holographic and speckle interferometry for vibration analysis are described. In the case of digital double-pulse holographic interferometry, two separate holograms of an object under test are recorded within a few microseconds using a CCD camera and stored in a frame grabber. The phases of the two reconstructed wave fields are calculated from the complex amplitudes which are obtained by digital reconstruction of the wavefront produced by the hologram. The deformation is obtained from the phase difference. In the case of ESPI (or image plane hologram) the phase can be calculated by using the sinusoid-fitting method or the Fourier method. Using three directions of illumination and one direction of observation, all the information necessary for the reconstruction of the three-dimensional deformation vector can be recorded at the same time. Applications of the method for measuring rotating objects are discussed, together with the derotator needed.     

基于光栅投影的工件表面缺陷检测方法研究

针对工件厚度与表面缺陷检测精度不足的问题,提出一种基于双投影的光栅投影缺陷检测方法。采用CCD相机采集两个方向上投影到工件表面的云纹图像,利用相移法求解两个方向的云纹图像相位,将两个方向的重建三维轮廓进行融合,实现工件表面缺陷的检测。试验结果证明:工件的厚度测量精度为0.05mm,缺陷测量精度提高到0.1mm。     

齿面干涉图像的相位解包裹算法对比研究

齿面干涉图像具有局部散斑、边缘杂散条纹、疏密不均和条纹粘连等噪声特征。为了寻找适用于处理齿面干涉图像噪声特征的解包裹算法,选取枝切法、相位导数偏差法的质量图引导法、最小二乘法、傅里叶变换轮廓术以及傅里叶变换的质量图引导法等 5 种算法,分析其对齿面干涉图像噪声处理的解包裹特点,通过对比分析解包裹相位图以及结果图中相位数据曲线,得出相位导数偏差的质量图引导法相比于其他算法解包裹精度高,能较好地抑制因噪声严重导致的解包裹相位突变、连续波动等问题。