Reconstruction of the Point-Spread Function of the Human Eye Using a New Phase-Retrieval Algorithm

The double-pass method is thought to obtain the point spread function (PSF) in human eyes based on two techniques: the symmetric double-pass method using the same pupil size and the asymmetric double-pass method using a different pupil size. The symmetric double-pass method provides autocorrelation of the retinal PSF and, thus, the modulation transfer function. The asymmetric double-pass method provides low-frequency partial-phase information and the partial phase-retrieval algorithm is applied to obtain the complete-phase information, and to estimate the PSF. The partial phase-retrieval algorithm is based on the iteration method proposed by Fineup and Kowalczyk and requires a lengthy computation. In this study, we propose a new high-speed phase-retrieval algorithm based on the property that the real and imaginary parts of optical transfer functions (OTFs) continuously change in value.     

自动调焦系统中一种调焦的判别方法

准确有效的调焦评价函数是采用数字图像处理技术实现自动调焦的关键环节,而调焦评价函数的好坏又直接影响着调焦系统的结构精度和运算的复杂程度。分析了灰度标准差比较法和灰度对比度比较法的优缺点,提出了经改进的灰度比较法作为调焦评价函数,该方法具有响应速度快、计算量小、信噪比高、可靠性好等特点。实验与分析结果表明,该方法对具有复杂细节和纹理特征的图像能得到较好的处理结果,而对于背景高亮度的图像则不能判断其是否聚焦。     

横向各向同性地层快速各向异性反演方法

斯通利波的频散特性常用于反演横向各向同性地层的各向异性。该文重点对斯通利波频散曲线随井孔和地层弹性模量相关参数的变化规律进行了探讨,提出了频散曲线的快速插值计算方法,经过误差分析验证了插值频散计算的可行性。据此提出了通过对参数大间隔取值建立理论频散数值表,插值计算不同各向异性参数的频散曲线,对比波形频散快速反演横向各向同性地层各向异性的方法。基于实轴积分法计算的横向各向同性地层井孔声场分别进行了插值频散快速反演和传统方法反演,对比结果证实了该文提出的快速反演方法的准确高效性,并且对数据加噪后验证了快速反演方法的抗噪性。     

A new volume-of-fluid method with a constructed distance function on general structured grids

A second-order volume-of-fluid method (VOF) is presented for interface tracking and sharp interface treatment on general structured grids. Central to the new method is a second-order distance function construction scheme on a general structured grid based on the reconstructed interface. A novel technique is developed for evaluating the interface normal vector using the distance function. With the normal vector, the interface is reconstructed from the volume fraction function via a piecewise linear interface calculation (PLIC) scheme on the computational domain. Several numerical tests are conducted to demonstrate the accuracy and efficiency of the present method. In general, the new VOF method is more efficient than both the high-order level set and the coupled level set and volume-of-fluid (CLSVOF) methods. The results from the new method are better than those from the benchmark VOF method, particularly in the under-resolved regions, and are comparable to those from the CLSVOF method. Breaking waves over a submerged bump and around a wedge-shaped bow are simulated to demonstrate the application of the new method and sharp interface treatment in a two-phase flow solver on curvilinear grids. The computational results are in good agreement with the available experimental measurements.     

基于深度聚类的目标细粒度分类方法北大核心CSCD

为提高光电系统对弱小目标的识别和分类能力,降低算法对硬件平台和数据的依赖,提出一种无监督分类方法−基于目标深度特征聚类的细粒度分类方法。该方法通过轮廓、颜色、对比度等浅层特征提取提示目标,经超分辨处理后,利用卷积神经网络对目标的深层特征进行编码,进一步采用基于注意机制的主成分分析方法进行降维生成表征矩阵,最后利用聚类的方式实现目标细粒度分类。实验验证了基于不同神经网络的深度聚类方法在不同数据集上的分类性能,其中采用ResNet-34聚类方法在CIFAR-10测试集上细粒度分类性能达92.71%,结果表明,基于深度聚类的目标细粒度方法能够取得与强监督学习方法相当的目标分类效果。此外,还可以根据不同簇数和聚类等级的选择实现不同细粒度的分类效果。     

改进联合多站多伯努利滤波器的水下多目标跟踪

针对水下多站多目标跟踪问题,基于多站多伯努利滤波器思想,提出改进联合多站多伯努利滤波器的水下多目标跟踪方法。该方法有两处改进:首先在目标数变化场景,为了获得新生目标状态,提出基于节点分对组合随机观测新目标生成方法,该方法将观测节点两两分对,并进行观测对间组合,以相同概率随机选择观测对组合来生成新目标信息;其次为了输出连续、稳定航迹,提出基于统计双门限的航迹生成方法,该方法对超过第一门限的检测次数进行累加,并将结果与第二门限比较来控制目标航迹输出。仿真结果表明:(1)基于节点分对组合随机观测新目标生成方法能够正确生成新目标信息,与全局交叉定位目标生成方法相比,算法复杂度降低,并与节点数成线性关系;(2)基于统计双门限航迹生成方法能够实时输出稳定航迹,并且其最优子模式指派(OSPA)距离对比不使用该航迹生成方法时减少20%左右。      

一种基于连续小波变换的LIBS光谱自动寻峰方法

谱线寻峰在激光诱导击穿光谱(laser-induced breakdown spectroscopy)中是必不可少的步骤,然而噪声和背景的存在,干扰了寻峰的准确性。为了提高LIBS寻峰中重叠峰分辨能力和自适应能力,提出一种适用于LIBS谱图的自动寻峰方法。将基于连续小波变换的脊线寻峰法引入到LIBS中,探讨了连续小波变换中的小波母函数的选择并优化了尺度参数和平移参数,提出一套脊线校正方法对脊线寻峰法进行了改进。实验结果表明,该方法与直接比较法、导数法和脊线寻峰法相比,在重叠峰分辨能力和寻峰精度等方面优势显著,可应用于LIBS数据处理。     

A Sound Field Separation and Reconstruction Technique Based on Reciprocity Theorem and Fourier Transform

We show a method to separate the sound field radiated by a signal source from the sound field radiated by noise sources and to reconstruct the sound field radiated by the signal source. The proposed method is based on reciprocity theorem and the Fourier transform. Both the sound field and its gradient on a measurement surface are needed in the method. Evanescent waves are considered in the method, which ensures a high resolution reconstruction in the near field region of the signal source when evanescent waves can be measured. A simulation is given to verify the method and the influence of measurement noise on the method is discussed.     

Numerical reconstruction stability of fiber Bragg gratings

Traditional algorithms for solving the inverse scattering problem for a fiber Bragg grating are described, and a new numerical method for this problem is developed. The method is based on the fast inversion of a matrix using its Toeplitz symmetry and on a special “inner-bordering” procedure. It is shown that the method is equally efficient as the well-known discrete layer peeling method but exceeds the latter in accuracy, especially for high-reflectance gratings. The stability of the proposed method with respect to initial data errors is demonstrated.     

Fresnel diffraction method with object wave rotation for numerical reconstruction of digital hologram on tilted plane

We present a new method for numerically reconstructing holography image on tilted planes within the Fresnel domain. The method is based on the Fresnel diffraction algorithm and object wave rotation transform. By suitable rotation the object wave on recording plane and once fast Fourier transformation this method can be implemented. The proposed method can render the real image of the hologram not only on parallel plane but also on the tilted. Experimental results are presented to demonstrate the method for a single axis rotation.     

Quantitative estimation of ultrasonic attenuation in a solid in the immersion case with correction of diffraction effects

This paper presents a method of diffraction correction for the log-spectral difference method to estimate quantitatively attenuation of a solid in the immersion case. The correction method is established based on the angular spectrum approach that is used to calculate the echoes from the front and back surfaces of the immersed solid. An example is given of a copper plate submerged in water and inspected by a linear array with a cylindrically curved surface. The correction method is first applied to a theoretical estimation of attenuation which is linearly dependent on frequency. The results have shown that the evaluated attenuation coefficient is in excellent agreement with the exact value. Then the method is applied to a real situation, in which the results have shown that the method yields reasonable evaluated attenuation values. This work has demonstrated that the method is able to correct effectively the diffraction effect so as to achieve a quantitative estimation of attenuation.     

Modular Fisher Discriminant Sparse Representation for robust face recognition

This paper proposes a novel framework for robust face recognition based on sparse representation and discrimination ranking. This method consists of three stages. The first stage partitions each training sample into some overlapped modules and then computes each module's Fisher ratio, respectively. The second stage selects modules which have higher Fisher ratios to comprise a template to filter training and test images. The dictionary is constructed by the filtered training images. The third stage computes the sparse representation of filtered test sample on the dictionary to perform identification. The advantages of the proposed method are listed as follows: the first stage can preserve the local structure. The second stage removes the modules that have little contribution for classification. Then the method uses the retaining modules to classify the test sample by SRC which makes the method robust. Compared with the related methods, experimental results on benchmark face databases verify the advancement of the proposed method. The proposed method not only has a high accuracy but also can be clearly interpreted.     

Scanning moiré and spatial-offset phase-stepping for surface inspection of structures

In order to develop a high-speed and accurate surface inspection system of structures such as tunnels, a new surface profile measurement method using linear array sensors is studied. The sinusoidal grating is projected on a structure surface. Then, the deformed grating is scanned by linear array sensors that move together with the grating projector. The phase of the grating is analyzed by a spatial offset phase-stepping method to perform accurate measurement. The surface profile measurements of the wall with bricks and the concrete surface of a structure are demonstrated using the proposed method. The change of geometry or fabric of structures and the defects on structure surfaces can be detected by the proposed method. It is expected that the surface profile inspection system of tunnels measuring from a running train can be constructed based on the proposed method.     

SPH simulation of fuel drop impact on heated surfaces

The interaction of liquid drops and heated surfaces is of great importance in many applications. This paper describes a numerical method, based on smoothed particle hydrodynamics (SPH), for simulating n-heptane drop impact on a heated surface. The SPH method uses numerical Lagrangian particles, which obey the laws of fluid dynamics, to describe the fluid flows. By incorporating the Peng–Robinson equation of state, the present SPH method can directly simulate both the liquid and vapor phases and the phase change process between them. The numerical method was validated by two experiments on drop impact on heated surfaces at low impact velocities. The numerical method was then used to predict drop-wall interactions at various temperatures and velocities. The model was able to predict the different outcomes, such as rebound, spread, splash, breakup, and the Leidenfrost phenomenon, consistent with the physical understanding.     

基于空时级联滤波的拖船自噪声抵消方法

拖船自噪声属于近场多途强干扰信号,具有显著角度扩展现象,是影响拖曳阵探测性能的重要因素之一。在逆波束形成自噪声抵消方法基础上,结合空域和时域滤波,提出了空时级联滤波的拖船自噪声抵消方法。首先,由方位估计器估计拖船自噪声方位,经过波束形成空域滤波估计得到自噪声信号;然后,利用自噪声空域滤波输出设计最优维纳滤波器,将自噪声抵消在阵元级。基于简正波声场模型的仿真和实际海试数据的处理表明,该方法能够有效抵消拖船自噪声,性能优于逆波束形成方法。     

元胞方法与蒙特卡洛方法相结合的薄膜生长过程模拟

利用仿真方法从原子尺度研究薄膜生长过程是当前薄膜研究领域的热点. 目前, 仿真方法主要在纳米尺度模型实现, 时空需求很大. 针对这一问题, 本文提出元胞和蒙特卡洛相结合的模拟方法, 实现对微米尺度模型薄膜生长过程的模拟. 利用元胞方法来实现模型表示以及演化计算, 从而降低对内存空间的要求, 提高计算效率, 并使用蒙特卡洛方法计算粒子的扩散概率. 通过对氮化硅薄膜生长过程进行具体研究, 将模拟结果与实际实验结果和分子动力学演化结果进行表面形貌和成分的比较, 验证了该方法的有效性.     

基于递阶模糊聚类的混沌时间序列预测

提出一种新的基于递阶模糊聚类系统的模糊建模方法.目的在于通过一系列的步骤优化T-S模糊模型结构，实现非线性系统的建模和预测.首先利用最近邻聚类法初始划分输入空间，得到规则数及初始聚类中心，用模糊C均值算法(FCM)进一步优化聚类中心；然后利用加权最小二乘法估计模糊模型的初始参数，进一步利用带遗忘因子的递推最小二乘法优化结论参数.采用该方法对Mackey-Glass混沌时间序列进行预测实验，结果表明可以对Mackey-Glass混沌时间序列进行准确建模和预测，证明了本方法的有效性. 关键词： 递阶模糊聚类 模糊建模 混沌时间序列 最小二乘     

Shape and topology optimization based on the phase field method and sensitivity analysis

This paper discusses a structural optimization method that optimizes shape and topology based on the phase field method. The proposed method has the same functional capabilities as a structural optimization method based on the level set method incorporating perimeter control functions. The advantage of the method is the simplicity of computation, since extra operations such as re-initialization of functions are not required. Structural shapes are represented by the phase field function defined in the design domain, and optimization of this function is performed by solving a time-dependent reaction diffusion equation. The artificial double well potential function used in the equation is derived from sensitivity analysis. The proposed method is applied to two-dimensional linear elastic and vibration optimization problems such as the minimum compliance problem, a compliant mechanism design problem and the eigenfrequency maximization problem. The numerical examples provided illustrate the convergence of the various objective functions and the effect that perimeter control has on the optimal configurations.     

The study of fractal correlation method in the displacement measurement and its application

The classical digital speckle, or digital image, correlation method of deformation measurement is based on gray level correlation between unformed and deformed digital images. The pattern of artificial random speckles and the natural textures on some object's surfaces have fractal characteristics, and their fractal dimensions represent both gray and morph information. Furthermore, the fractal dimensions are stable feature parameters of the patterns. The digitized images of the patterns are confirmed to be also fractals. By this fact, a new method of displacement measurement is developed in the paper, based on the fractal dimensions correlation. The in-plane displacement fields of a body can be acquired. In order to verify the validity of the new method, an experiment has been designed and the results have been compared with those obtained from the classical digital image correlation method. The validity of the new method is not less than that of classical method. Further discussions about the traits and the developing vista of the method are given at the end.     

Computation of localized modes in a defect-containing photonic crystal by the method of periodic continuation of solution

The applicability of the method of periodic continuation of solution for computing the spatial distribution of an electromagnetic field in a photonic crystal with a defect is analyzed. The accuracy of the method is estimated. The crystal eigenmodes are classified, and the method is shown to be applicable only for states localized on the defect. The results of numerical calculations based on expansion of electromagnetic field in plane waves are presented, and the details of the expansion method are described.