相机位姿估计的加速正交迭代算法

面对需要实时计算的相机位姿估计问题,针对经典的广泛应用的正交迭代算法,提出了一种加速正交迭代算法。其关键思想是将每一次迭代过程规整化,从而提炼出每一次迭代的重复计算,若将此重复计算在迭代开始前提前计算,则可以大幅度的减少迭代过程中的计算量,使得每一次迭代的计算复杂度从O(n)降低为O(1)。因此,可以在更短的时间内迭代更多的次数,从而获得更高的精度。进行了对比实验,结果显示本加速算法计算精度更高,速度更快。并通过实验提出了选择稳健n点透视(RPn P)计算初值,再使用加速正交迭代算法进行迭代运算的方法,在控制点不多的情况下,是一种精度接近最大似然估计,计算速度最快的算法。     

基于GPU求解椭圆型偏微分方程的并行算法

针对求解椭圆型偏微分方程的雅克比迭代算法和DRM算法进行基于GPU的CUDA加速算法研究.通过两个算例在GTX570显卡上对GPU加速算法进行验证.结果表明,在保证运算精度的前提下,雅克比迭代的GPU加速效率最高,在DOUBLE类型下的加速比可达到14倍左右,效率可达到53%左右;DRM算法在DOUBLE类型下的加速比最高可达到3.8倍,效率达到15%左右.     

基于似然函数最速下降的红外与可见光图像配准

为了实现红外与可见光图像的自动配准,提出了基于似然函数最速下降迭代的图像配准算法.该算法以图像边缘作为配准点特征,将异源图像配准转化为边缘点集配准.基于点集的高斯混合模型建立了边缘点集配准似然函数,以该函数作为目标函数,仿射变换参量作为优化变量,利用最速下降方法进行最优变换参量求解,从而实现边缘点集配准.同时,将多分辨率金字塔引入迭代配准框架下,实现了高分辨率图像配准的加速.实验结果表明:该算法精度高,运算速度快,可以很好地完成可见光与红外图像的自动配准.     

扫描相干衍射成像背景噪声分离算法研究

X射线扫描相干衍射成像(ptychography)是一种新型的无透镜成像方法,摆脱了传统透镜成像中聚焦元件对分辨率的限制,使理论分辨率只受到X射线波长和探测器数值孔径的限制.然而实验测量中的噪声限制了该方法对成像质量的改善,甚至最终导致图像重建失败.在研究了ptychography现有的相位恢复迭代算法后,本文提出了一种新型的图像重建迭代算法.该算法利用ptychography数据的高冗余性,通过梯度下降最小化技术,在重建样品和探针图像的同时还完成了背景噪声的同步迭代重建,实现了信号和噪声的盲分离功能.通过仿真模拟和实验数据重建,将该方法与传统的迭代算法进行了对比,结果表明新算法能够较好地实现信噪分离,显著提升ptychography的成像质量.     

利用加权预测的图像迭代盲解卷积

为解决大气湍流造成的图像退化问题,本文鉴于现有的盲解卷积算法收敛性不稳定,计算量大等特点,提出了一种基于加权预测的迭代盲解卷积算法。对目前性能优秀的用迭代实现盲解卷积的L-R算法进行优化,在每次迭代结束后通过加权方法求出预测值,根据预测值计算方向加速算子,从而大大提高算法的收敛速度。实验表明：该算法不仅可对模糊退化图像进行很好的复原,同时与L-R算法相比收敛速度提高约43.8倍,其迭代速度快的特点决定了算法具有较高的工程实用价值。     

利用加权预测的图像迭代盲解卷积

为解决大气湍流造成的图像退化问题,本文鉴于现有的盲解卷积算法收敛性不稳定,计算量大等特点,提出了一种基于加权预测的迭代盲解卷积算法。对目前性能优秀的用迭代实现盲解卷积的L-R算法进行优化,在每次迭代结束后通过加权方法求出预测值,根据预测值计算方向加速算子,从而大大提高算法的收敛速度。实验表明:该算法不仅可对模糊退化图像进行很好的复原,同时与L-R算法相比收敛速度提高约43.8倍,其迭代速度快的特点决定了算法具有较高的工程实用价值。     

基于Zernike模式的随机并行梯度下降算法的收敛速率

为了加快控制变形镜进行波前整形的随机并行梯度下降（SPGD）算法的收敛速率,提高实时波前整形能力,本文利用由12阶Zernike多项式构成的畸变波前和32单元变形镜建立了仿真模型。基于Zernike多项式的单位正交性,得到了两个常数矩阵,当斯特列尔比（SR）达到0.8时,需要算法迭代660次,简化了算法的运算过程,加快了算法运行时间。通过Matlab7.8.0对6种SPGD算法进行仿真对比,结果显示：当SR要求不高时,可使用间接固定双边SPGD算法来提高收敛速度;当SR要求较高时,则应当使用间接自动双边SPGD算法。提出的算法为实际的激光整形提供了理论指导。     

扩展目标夏克-哈特曼波前传感器的频域迭代算法研究

扩展目标夏克-哈特曼波前传感器(SH-WFS)子图像之间偏移量的计算是影响波前传感精度的关键,通常采用相关算法来实现,并通过抛物线插值达到亚像元精度。子图像间的相对偏移量计算也可采用频域相移量估算的方法进行计算,频域算法还可通过迭代进一步提高算法精度。对频域迭代算法进行了理论分析、仿真和实验研究,结果表明,频域迭代算法在信噪比高于4…1时,具有比抛物线插值法更高的计算精度;在信噪比较低时,与抛物线插值法精度相近。     

解卷积的多重信号分类算法方位谱低背景处理方法

针对信噪比较低时,多重信号分类(Multiple Signal Classification,MUSIC)算法方位谱背景级较高的问题,提出了一种解卷积的MUSIC方位估计算法(Deconvolvecd MUSIC,D-MUSIC)。该方法用一个类似冲激函数作为MUSIC算法输出方位谱的点散射函数(Point Scattering Function,PSF),然后基于解卷积图像复原理论,利用该点散射函数和RichardsonLucy(R-L)迭代算法对MUSIC算法的方位谱进行解卷积,获得D-MUSIC算法的方位谱,达到降低方位谱背景级的目的。仿真表明,该方法继承了MUSIC算法的高分辨性能,且可以明显降低方位谱的背景级,具有较好的方位估计性能。对南海海上试验的水平阵数据进行处理,分析比较了利用MUSIC算法和解卷积MUSIC算法获得的方位谱时间历程图,分析结果有效验证了D-MUSIC算法性能的优越性。     

面向光子网格任务调度的迭代列表算法

光子网格中任务和通信的联合调度是一个非确定性多项式难题.为了进一步优化调度长度,本文在扩展列表算法的基础上,提出一种迭代列表调度算法.该算法通过扩展列表算法产生一个初始调度序列,并通过迭代的方式不断估计调度过程中子任务之间的通信时间;然后重新计算子任务的权重,调整子任务的调度序列,达到改善调度长度的目的.仿真实验表明,迭代调度算法对于大部分的实例能够有效地减少任务的调度长度,并且更加适用于数据密集型的任务调度.     

Application of the vector ε and ρ extrapolation methods in the acceleration of the Richardson-Lucy algorithm

The vector ε and ρ extrapolation methods are applied in accelerating the convergence of the Richardson-Lucy (R-L) algorithm and its damped version. The theory and implementation are discussed in detail, and relevant numerical results are given, including the cases of noise-free images and images corrupted by the Poisson noise. The results show that the vector ε and ρ extrapolations of 9 orders can speed the convergence quite efficiently, and the ρ(9) method is more powerful than the ε(9) method for noisy degraded images. The extra computation burden due to the extrapolation is limited, and is well paid back by the accelerated convergence. The performances of these two methods are compared with the famous automatic acceleration method. For noise-free degraded images, the vector ε(9) and ρ(9) methods are more stable than the automatic method. For noisy degraded images, the damped R-L algorithm accelerated by vector ρ(9) or automatic methods is more powerful, and the instability of the automatic method is restrained by the damping strategy. We explain the instability of the method in accelerating the normal R-L algorithm by the numerical noise due to its frequent applications in the run.     

Delay-Doppler deconvolution image formation for multiple targets in waveguide environment

The ranges and radial velocities of targets are interesting parameters in the active sonar, which can't be observed directly but can only be estimated. Firstly, by making use of the delay-Doppler model of multi-targets in waveguide multipath environment, one finds that sample cross-ambiguity function is a two-dimensional(2D) convolution of the auto-ambiguity function of the transmitted signal with the generalized target reflectivity density, which is a 2D convolution of the spread function of channel with the reflectivity density as well. Secondly,from the perspective of information theory, an iterative deconvolution algorithm named R-L(Richardson-Lucy) is derived based on minimum Csiszar discrimination criterion. Finally, the blurs caused by both of the transmitted signal and channel are removed by 2D deconvolution of the expectation of sample cross-ambiguity function, 2D image and then parameters of time-delay and Doppler is estimated sequentially. Results of both numerical simulation and sea experimental data processing verify the feasibility and effectiveness of R-L deconvolution algorithm, which effectively improves the resolution and precision of the time delay and Doppler estimation, when compared to the classical match filtering and Wiener filtering.     

波导环境多目标时延-多普勒解卷成像

有源声呐感兴趣的参量是目标距离和径向速度,它们无法直接观测得到,需要通过估计而获得。利用波导多路径环境多目标时延-多普勒模型,可以导出采样互模糊度函数均值是发射信号自模糊度函数与广义目标反射性密度函数的两维卷积,其中广义目标反射性密度函数为信道扩展函数与目标反射性密度函数的两维卷积。依据信息理论最小Csiszar鉴别准则,可导出R-L (Richardson-Lucy)迭代解卷算法,对采样互模糊度函数均值进行两维迭代解卷积,消除发射信号和信道引入的模糊,序贯地实现时延-多普勒两维像的估计,进而获得多目标的时延和多普勒参量估计。仿真结果和海上实验数据分析验证了R-L解卷算法的可行性和有效性,较之常规的匹配滤波和维纳滤波算法,R-L算法有效地提高了时延和多普勒估计的分辨力和精度。     

Noise-Robust Image Reconstruction Based on Minimizing Extended Class of Power-Divergence Measures

The problem of tomographic image reconstruction can be reduced to an optimization problem of finding unknown pixel values subject to minimizing the difference between the measured and forward projections. Iterative image reconstruction algorithms provide significant improvements over transform methods in computed tomography. In this paper, we present an extended class of power-divergence measures (PDMs), which includes a large set of distance and relative entropy measures, and propose an iterative reconstruction algorithm based on the extended PDM (EPDM) as an objective function for the optimization strategy. For this purpose, we introduce a system of nonlinear differential equations whose Lyapunov function is equivalent to the EPDM. Then, we derive an iterative formula by multiplicative discretization of the continuous-time system. Since the parameterized EPDM family includes the Kullback–Leibler divergence, the resulting iterative algorithm is a natural extension of the maximum-likelihood expectation-maximization (MLEM) method. We conducted image reconstruction experiments using noisy projection data and found that the proposed algorithm outperformed MLEM and could reconstruct high-quality images that were robust to measured noise by properly selecting parameters.     

基于模糊松弛迭代的分层图像增强算法

映射函数的研究是图像增强技术的关键,模糊松弛技术利用模糊函数建立其映射函数,通过设置该算法的模糊松弛参数可以达到对图像可控增强处理的目的。为了实现针对不同层次的图像内容实现可控式模糊增强,提出了多层次交互迭代的模糊松弛图像增强算法。通过分析各子层图像的直方图,分别设定各子层的模糊松弛参数,并建立对应的模糊松弛映射函数;分别对各子层图像进行可控迭代增强,选择合适的迭代次数作为终止条件。给出的实验结果证明了该算法能够取得较好的增强效果。     

一种全局参数估计的水下主动偏振去雾算法

为改善水下主动光照明条件下后向散射光对成像的影响,通过分析水下主动偏振成像模型,提出一种基于拟合函数的全局参数估计的水下主动偏振去雾算法。该算法结合图像增强作为图像预处理,再设定二元多项式函数,利用最小二乘法进行后向散射光偏振度变量的拟合求解,得到对比度更高、信息更丰富的水下复原图像。实验结果表明:该算法可有效改善主动光照明条件下的水下图像质量,提高图像对比度,复原被淹没的细节信息,复原图像的图像增强测量值较以往算法相比提升70%,且能适用于不同浓度介质的情况。     

非精确交替方向总变分最小化重建算法

CT (computed tomography)系统实际应用当中, 经常会出现扫描数据不满足数据完备性条件的情况. 针对不完全角度重建问题的研究, 是目前迭代型算法研究中的一个热点. 一系列基于带有约束的总变分最小化的重建算法近年来在不完全角度重建中取得了较好的效果, 这其中基于交替方向法 (alternating direction method, ADM) 的重建算法表现出更好的性能. 然而, ADM方法在求解过程中对矩阵求逆的处理效率不高, 导致极大的计算开销. 本文针对该问题, 使用非精确ADM方法, 利用线性近似的方式替换掉计算开销较大的项, 使得矩阵求逆问题可以通过快速傅里叶变换加速实现. 实验结果表明, 本文提出的非精确交替方向总变分最小化重建算法与精确ADM重建算法相比, 没有明显的精度损失, 计算时间缩减30%左右. 关键词： 不完全角度重建 总变分最小化 非精确交替方向法     

Digital image correlation using iterative least squares and pointwise least squares for displacement field and strain field measurements

Digital image correlation (DIC) method using iterative least squares algorithm (ILS) for displacement field measurement and pointwise least squares algorithm (PLS) for strain field measurement is proposed in this paper. A more general and practical intensity change model is employed with consideration of the linear intensity change of the deformed image, followed by an iterative least squares algorithm for calculating displacement field with sub-pixel accuracy. The concept of correlation function is not used in the ILS method, even though we prove that the algorithm is actually equivalent to the optimization of the sum of squared difference correlation function using improved Newton–Raphson method. Besides, different from the conventional strain estimation approaches based on smoothing the displacement fields first and followed by differentiation of the smoothed displacement fields, a simple yet effective PLS algorithm is proposed for extracting strain fields from the computed displacement fields. The effectiveness and accuracy of the proposed techniques is verified through numerical simulation experiments. A practical application of the algorithms to residual plastic deformation field measurement of GH4169 alloy subjected to tensile fatigue is also presented.     

Comparative study on several blind deconvolution algorithms applied to underwater image restoration

Although the use of blind deconvolution of image restoration is a widely known concept, only few reports have discussed in detail its application to solving problem of restoration of underwater range-gated laser images. A comparative study of underwater image restoration using the Richardson-Lucy algorithm, the least-squares algorithm, and the multiplicative iterative algorithm for blind deconvolution is presented. All the deconvolution approaches use denoised underwater images and Wells’ small angle approximation theory of derived point spread function as the initial object and degradation guess, respectively. Owing the underwater no-reference imaging environment, image quality judgment based on the blur metric method is incorporated in our comparison to determine the appropriate deconvolution iteration number for each algorithm, which objectively evaluates the image restoration results. The performance of the three algorithms applied to underwater image restoration is discussed and reported.     

保留边界特征的深度图像增强算法研究

针对现有深度图像增强算法存在边界保留特性差的问题,提出梯度掩模导向联合滤波(gradient mask guided joint filter, GMGJF)算法。利用深度图像进行Sobel梯度变换获取边界方向信息,利用深度图像空洞区域生成空洞掩模,再以边界方向和空洞掩模为导向联合彩色图像对深度图像进行迭代高斯滤波和空洞填充。实验结果表明,GMGJF算法的PSNR(peak signal to noise ratio)、SSIM(structural similarity index measure)比IMF(iterative median filter)、GF(guided filter)、JBF(joint bilateral filter)算法的PSNR、SSIM至少提高了3.50%和1.07%,不仅去噪能力、空洞填充能力最强,而且边界特征保持最好,有利于深度图像的特征提取与目标识别。