小波变换轮廓术中用小波脊系数幅值引导相位展开的研究

为了减少小波变换轮廓术中相位展开过程的误差传递,从小波变换的相关实质出发,提出了在小波变换轮廓术中利用以前被忽略的小波脊系数幅值作为可靠度判据指导相位展开的方法.该方法选择最大幅值的小波脊系数所在位置作为相位展开起始点,根据小波脊系数幅值的大小,确定一条由大幅值到小幅值的最优化的展开路径,最大限度减少了相位展开过程中的误差传递.由于充分利用了小波变换系数的幅值信息,最大限度减小了相位展开过程中的误差传递.计算机模拟和实验验证了基于小波脊系数幅值的相位展开方法的正确性.     

用于相位展开的调制度-罗伯特梯度幅值方差质量图

为了减少相位展开过程中噪音的影响并获得正确的相位展开结果,提出了一种新型的质量图。首先分析了传统的相位导数方差质量图和基于调制度信息质量图的缺陷。然后基于各向同性的梯度幅值,融合图像锐化方法和调制度信息,构造出一种新型的"调制度-罗伯特梯度幅值方差"质量图。最后通过计算机模拟图和模型实例图的相位展开,将提出的质量图与传统的相位导数方差和基于调制度信息的质量图进行了对比。实验结果表明,新的质量图更加可靠,有效地弥补了相位导数方差等传统质量图的部分缺陷。     

基于图切割的相位展开

剪切散斑干涉术中,普遍采取相移法提取相位,只可得到被包裹的相位信息,进一步量化计算之前,必须展开相位,但是大量的散斑噪音使相位展开变得非常困难.本文提出基于图切割理论的相位展开算法,将相位展开等价于整数的最优估计问题,通过最小化能量函数展开相位,无需对散斑包裹相位图进行滤波,就可以从包含大量散斑噪音的包裹相位图中准确地提取出真实相位,极大程度地保留了相位包裹图中的细节信息.对于同一幅模拟的散斑包裹相位图,传统的质量导向枝切法和最小二乘法的均方根误差分别为11.707 6和4.977 5,新算法的均方根误差则为0.945 9,数值模拟与实验结果均验证了新算法优良的抗噪性能.     

基于图切割的相位展开

剪切散斑干涉术中,普遍采取相移法提取相位,只可得到被包裹的相位信息,进一步量化计算之前,必须展开相位,但是大量的散斑噪音使相位展开变得非常困难.本文提出基于图切割理论的相位展开算法,将相位展开等价于整数的最优估计问题,通过最小化能量函数展开相位,无需对散斑包裹相位图进行滤波,就可以从包含大量散斑噪音的包裹相位图中准确地提取出真实相位,极大程度地保留了相位包裹图中的细节信息.对于同一幅模拟的散斑包裹相位图,传统的质量导向枝切法和最小二乘法的均方根误差分别为11.7076和4.9775,新算法的均方根误差则为0.9459,数值模拟与实验结果均验证了新算法优良的抗噪性能.     

基于指数尺度间隔连续小波变换的相位提取算法

只需要一幅调制图像的光栅投影测量方法主要有傅里叶变换轮廓术(FTP)、小波变换轮廓术(WTP)等。采用基于指数尺度间隔的连续小波变换与重构方法,提取调制图像的瞬时相位。针对指数尺度间隔连续小波变换,指出了足够大的噪声能够改变小波变换脊的位置,并且该脊向上移动的概率最大。因此,为了重构载频信号,选择脊及其紧邻的较大的那个尺度所对应的小波系数,采用灰度图像阈值分割中最大类间方差法(OTSU),剔除掉幅值较小的系数;针对斑点噪声的影响,对OTSU算法的结果进行了修正;使用修正后的系数集合重构载频信号,并计算该信号的瞬时相位。理论分析和实验结果表明算法有效且具有稳健性。     

非单调条纹图的相位恢复新方法

在连续小波变换和展开相位拐点识别的基础上,提出一种非单调条纹图的相位恢复新方法.包裹相位通过连续小波变换的方法提取,提出小波变换尺度步长的选择准则及对粗糙尺度作迭代均值滤波的技术.基于展开相位拐点与符号歧义点的单应性,可通过检测并修正展开相化拐点的方法恢复非单调条纹图的真实相位.给出了详细的理论推导、数值模拟及实验验证过程.数值模拟包括一维及二维的含噪声信号,实验为基于显微干涉法的微桥的静动态形貌和变形测试.模拟及实验结果表明,该法只需一幅条纹图即可准确解调非单调条纹的相位,最大误差小于4%,且具有极强的抗噪能力.这为显微干涉测量提供了有效的相位恢复新手段.     

局部噪声斑块InSAR干涉图相位解缠算法

针对干涉合成孔径雷达干涉图由于局部噪声、阴影或条纹断裂等原因导致残差点集中、产生不可靠数据点噪声斑块,从而引起相位解缠困难的情况,提出基于掩膜滤波的质量图引导权值的四向加权最小二乘相位解缠迭代算法.首先对噪声斑块区域进行掩膜滤波处理,有效保留该区域真实相位信息,防止大误差传播.再结合最大相位梯度和相位导数变化的质量图法来定义权值,构造新的四项最小二乘迭代法,并用其对掩膜区域进行填充,抑制误差传递、提高运算速度、改善过度平滑作用.模拟及真实数据实验表明,该方法解缠精度高,能真实还原干涉合成孔径雷达干涉图局部噪声斑块的原始相位.与传统最小二乘法相比,其差分方程矩阵带宽变大,填充效果更好,迭代次数更少,精确度更高,解决了因噪声集中导致干涉合成孔径雷达图像解缠难度大的问题.     

基于整数递推GCV的遥感图像去噪算法

本文提出了一种整数递推广义交叉校验算法.在对遥感图像做小波变换后,统计子带中小波系数的幅值分布,进行同值小波系数合并运算,然后通过整数优化递推加速广义交叉校验计算过程,减少相邻阈值下广义交叉校验函数的冗余计算.在对多幅遥感图像仿真实验中,当噪音标准差为10～30时,整数递推-广义交叉校验算法耗时仅为广义交叉校验算法的2%至0.5%,且保证了去噪效果一致,能有效提高图像PSNR0.66 dB至6.03 dB.此外,广义交叉校验算法耗时随噪音增大及遥感图像尺寸增长而迅速升高,整数递推-广义交叉校验算法耗时则相对平稳.     

基于多尺度总体最小二乘的图像去噪

提出了一种基于多尺度总体最小二乘的图像去噪算法.采用平稳小波变换对噪音图像进行分解,分别对各个分解层的高频子带,通过总体最小二乘算法估计信号小波系数;并且考虑到不同尺度小波系数之间的相关性,将尺度相关性约束到总体最小二乘算法中,进而准确估计各高频子带信号小波系数,再由估计的信号小波系数通过小波逆变换得到去噪图像.实验结果表明,考虑尺度间相关性的总体最小二乘平稳小波变换图像去噪算法能有效去除图像噪音,在信噪比和视觉质量上有了较大改善.     

基于多尺度总体最小二乘的图像去噪

提出了一种基于多尺度总体最小二乘的图像去噪算法.采用平稳小波变换对噪音图像进行分解,分别对各个分解层的高频子带,通过总体最小二乘算法估计信号小波系数|并且考虑到不同尺度小波系数之间的相关性,将尺度相关性约束到总体最小二乘算法中,进而准确估计各高频子带信号小波系数,再由估计的信号小波系数通过小波逆变换得到去噪图像.实验结果表明,考虑尺度间相关性的总体最小二乘平稳小波变换图像去噪算法能有效去除图像噪音,在信噪比和视觉质量上有了较大改善.     

Simultaneous spectral calibration of two phase plates

Based on a new model for the spectral calibration of phase plates in pairs, experimental results are presented for three mica plates between 430 and 684 nm. One of the plates is 36.4 μm thick and the remaining two are identical, each being 28.6 μm thick. In the studied spectral range, the retardances varied from 123.5° to 86.8° for the first plate and from 96.9° to 68° for each of the two other plates. Different procedures for performing the calibration of the plates and the special case of two identical plates are discussed.     

Influence of Cd addition on the phase transformations of β Cu–Al alloys

Cu–Al alloys in the hyper- and hypoeutectoid region with addition of Cd are studied. The phase transformation sequence during continuous heating and cooling has been followed by resistometry, calorimetry, and optical microscopy.     

条纹图象的数字化自动分析处理技术之二:相位分析法

相位分析法是目前光学条纹图象数字化自动分析处理中的最重要方法之一。本文较全面地介绍了相位分析技术,并对相位分析的关键技术--相位的滤波和解包络方法做了重点说明。文中给出了相位解包络的实例并讨论了图象数字化自动分析处理技术的实际应用状况。     

Phase retrieval in two-wavelength DSSI using a combined filtering method

Phase retrieval in two-wavelength DSSI (digital speckle shearing interferometry) using a combined filtering method is proposed for small deformation derivative measurement. A simultaneous two-wavelength illumination and 3-CCD camera are employed in the experimental setup. The proposed method, which uses a two-wavelength technique does not require the conventional spatial phase unwrapping and has the advantage of good noise suppression for phase retrieval. Experimental results obtained show that using the proposed combined filtering method sensitivity similar to the single wavelength technique can be achieved.     

普通显微镜改装成相衬显微镜的一种新方法

本文提出了一种新的相衬显微镜的光路系统,这种系统非常适用于将普通显微镜改装成相衬显微镜。首先,叙述了相衬法的基本原理;接着提出了一种将相衬板置于物镜前方的相衬显微镜系统,并运用傅里叶光学证明了在非柯勒照明条件下在物镜前方确实存在物体的频谱面。这种相衬显微镜系统对于将普通显微镜改装成相衬显微镜来说,不仅简单实用,而且杂散光小,有利于提高灵敏度。     

异核多脉冲实验的相位循环设计方法

本文介绍了异核多脉冲实验相位循环的设计方法.先分别设计两个不同核的相位循环编码行列,然后根据正交检测相位要求把两个列数相等的行列以并联方式结合起来,就可得异核多脉冲实验的相位循环.     

光折变相位型波带片

利用成象方法和光折变效应,在LiNbO₃:Fe晶体中成功地写入了光折变相位型波带片.其衍射效率可达到21%.该制作方法具有实时、简便等优点.在集成光学和光学信息处理中具有重要的应用前景.     

Measurement of sub-surface delaminations in carbon fibre composites using high-speed phase-shifted speckle interferometry and temporal phase unwrapping

A high-speed phase-shifted speckle interferometer has been developed recently for studying dynamic events. Speckle interferograms are continuously recorded by a CCD camera operating at 1 kHz with temporal phase shifting carried out by a Pockels cell running at the same frequency. Temporal phase unwrapping through sequences of more than 1000 frames allows the determination of time-varying absolute displacement maps. This paper presents the application of this speckle interferometry system to the detection and measurement of sub-surface delamination defects in carbon fibre specimens. The influence of re-referencing the temporal phase unwrapping algorithm after different time intervals is analysed to reduce the random phase errors produced by speckle decorrelation and vibration. The performance of several phase-shifting algorithms to minimize the influence of the vibration noise caused by the vacuum pump used to load the specimen is also investigated.     

Phase structures and stability in barium titanate ferroelectric ultrathin films

Under taking into account both stress and surface effects, an extended Landau–Devonshire theory is developed to investigate the phase structures and ferroelectric stability in barium titanate ultrathin films. The phase diagrams of temperature-stress, film thickness-stress and polarization-film thickness are calculated. For the surface polarization smaller than the bulk, the paraelectric phase starts at temperature below the bulk Curie temperature under small stresses and the ferroelectric phases can be stabilized in the ultrathin films under large stresses. For the surface polarization larger than the bulk, the paraelectric phase appears at temperature above the bulk Curie temperature and the ferroelectricity can be sustained in the ultrathin films by the “superpolarized” surface layer.     

In situ calibration of a Michelson type, speckle-shearing interferometer: Wobbling mirror effect

In situ calibration of piezoelectric transducers in a phase stepping, Michelson type, speckle-shearing interferometer can be affected by mirror wobbling introduced by the transducers. When this happens, a modulation of the curves employed for calibration can be readily detected in imaging systems with large apertures. A calibration method that takes into account the modulation of these curves is, therefore, essential to insure the required regularity of the phase steps. A computer model of speckle-shearing interferometry is used to deal with piezoelectric non-linearities and mirror wobbling in the interferometer, and to show how the calibration curves are affected when both phenomena are present. Computer simulated and experimental results validate the calibration technique advanced here.