多通道图像色彩重建的神经网络算法

为了解决目前多光谱成像设备输出的应用问题,将神经网络算法应用于多通道图像色彩重建。CCD相机加LED光源组成多光谱成像系统,研究其输出的多通道图像中色彩信息的高效转换方法,建立了D50光源下的神经网络转换模型,利用BP神经网络和GRNN对多光谱成像系统进行色彩重建实验。研究结果表明：采用搭建的模型可以得到较高的色度重建精度,更换目标样本色彩重建性能保持稳定。     

基于维纳估计的光谱反射率重建优化算法研究

目的研究光谱反射率重建算法,解决各种物体颜色的光谱反射率重建精度问题。方法通过高精度多光谱成像系统获取实验样本的系统响应值,分光光度计获取样本的光谱反射率,采用Wiener估计法、自适应维纳估计法和提出的优化维纳估计法,对待测样本实验数据进行光谱重建,并评价重建结果。结果在3种光谱重建算法仿真实验中,提出算法的均方根误差平均值为0.0355,平均CIE1976色差为1.4349,优于其他2种算法。结论在光谱重建算法的研究中,基于优化的维纳估计算法可以有效提高光谱的重建精度,可应用于实际的多光谱成像复制中。     

基于改进白骨顶鸡优化算法的ECT图像重建北大核心CSCD

电容层析成像(ECT)技术求解图像重建问题存在严重的不适定性。为提高图像重建精度,提出了一种改进白骨顶鸡优化算法的ECT图像重建算法。针对标准白骨顶鸡优化算法(COA)中初始种群随机生成而导致的算法稳定性差和容易陷入局部最优的问题,引入了佳点集和正余弦优化方法,并融合了ART重建算法,根据ECT成像特点改进了目标函数。最后,进行仿真和静态实验,并与线性反投影(LBP)算法、Landweber算法、ART算法进行对比。结果表明,该方法可有效提高图像重建精度。     

基于图像的启发式三维火焰重建算法

研究了火焰的重建,为改进现有稀疏视角输入下的层析成像算法的重建精度,提出了一种基于图像的启发式火焰重建方法。该重建方法通过迭代优化求解能量约束模型,实现对火焰三维温度场的重建。算法允许复杂的成像模型应用于重建过程中,使渲染过程更接近火焰真实成像过程,并对数据平滑性进行约束,以提高重建结果质量。算法在迭代过程中使用启发式规则引导温度场数据调整,使算法快速收敛,并通过使用图形处理器加速重建过程。使用模拟数据和真实捕获数据进行实验,结果表明,与现有方法相比,提出的方法有效提高了重建结果的精度及平滑性,验证了提出方法的有效性。     

基于FCM优化的平面阵列电容成像算法

为提高平面阵列电容成像系统的成像精度,提出了一种基于FCM数据优化的成像算法。根据平面阵列电极电容数据的特点,为减小电容测量误差对介电常数的影响,利用FCM算法对测量电容值的不断收敛以实现数据优化的作用,在减弱噪声的同时提高电容值数据的稳定性。在此基础上,对一种隔热材料胶层进行缺陷检测实验,使重建图像的相关系数得到了提高,减小了图像重建误差。实验结果表明:图像重建结果的优化算法可获得更加稳定、有效的电容数据,胶层缺陷图像重建精度具有较大提升。     

一种电容层析成像图像重建优化算法

提出一种电学层析成像(ECT)图像重建优化算法。通过将传统正则化算法转化为最小二乘问题进行求解,结合lp范数逼近正则化最小化问题,利用重新加权的方法进行迭代计算。以油-气两相流模型进行仿真及静态实验,将所提出的优化算法与常用的LBP、Landweber迭代及Tikhonov正则化算法进行对比。结果表明,与常用算法相比,采用该优化算法对管道中心物体及多物体分布流型进行图像重建,其图像相对误差均为最低,且重建图像的形状保真度明显提高。     

ERT技术图像重建算法研究

两相流参数测量中,电阻层析成像技术的图像重建算法可归为两大类,即直接法和迭代法。对常用的几种图像重建算法进行了比较性研究,结果表明:直接法速度快,但重建图像质量不理想,仅可作定性分析;迭代法总体上说重建图像质量较高,但速度较慢。当增加像素数以提高重建图像分辨率时,直接法重建图像耗时增量较小,迭代型算法中改进的Newton-Raphson算法、BFGS算法由于目标函数的Hessian矩阵维数增加,计算时间大大增加,而改进的BFGS算法因简化了Hessian矩阵及其逆矩阵的计算,时间增量则较小。     

具有多约束连续体结构仿生拓扑优化方法

通过浮动参考区间法分析具有多约束连续体结构拓扑优化问题。浮动区间法是指将结构的拓扑优化过程看作是骨骼重建过程,通过引入参考应变区间,将结构中所有各点处主应变绝对值落入参考应变区间作为重建平衡状态,当结构处于重建平衡状态时获得结构的最优材料分布。为了使得优化结果满足给定的性态约束,参考应变区间在优化迭代过程中须不断变化。讨论了几种常见性态约束对结构性能的要求。给出了结构具有多约束时优化问题的算法。数值算例表明该方法可行。     

一种基于模型的光学层析图像重建方法

光学层析图像重建是个病态问题,测量误差会在重建过程中被放大,对此,提出一种以广义高斯马尔可夫随机场模型为先验信息的光学层析图像重建方法.重建过程是对目标函数的优化过程,目标函数关于光学参数的梯度计算是算法中的难点,因此,提出一种基于梯度树的梯度计算方法.文中分别给出了吸收系数和散射系数的重建结果,并引入三个指标因子衡量重建图像的质量,进而列出不同重建算法下,重建图像的指标值.最后通过对重建结果和指标因子取值的比较,分析基于模型的重建算法的有效性.     

基于主动位移成像的图像超分辨率重建

超分辨重建算法是一种将低分辨率图像恢复为高分辨率图像的算法,被广泛用于医学、遥感、军事安防以及人脸识别等领域。在黑夜、远场场景下构建数据集比较困难,基于深度学习的超分辨重建算法应用受到阻碍。而微扫描成像技术扫描模式固定,对器件到位精度要求高。针对这两个问题,我们提出一种基于主动位移成像的图像超分辨率重建算法。具体地,在控制相机随机移动的同时记录采样时刻位移,通过解算、映射选图、精确匹配图像序列并获取多帧图像间的亚像素信息,然后对估计图像进行迭代和更新,最后重建获得高分辨率图像。实验结果表明,本算法在PSNR、SSIM和平均梯度三个指标上都优于最近提出的基于POCS的图像超分辨率重建算法MFPOCS,与基于CNN的方法 ACNet相比具有竞争力。值得提出的是,本算法无需固定的扫描模式,降低了微扫描技术对器件实时到位精度的要求,同时,本算法可以保证重建初始帧的优良选取,有效规避了POCS算法的固有缺点。     

Multicriterion cross-entropy minimization approach to positron emission tomographic imaging

A multicriterion cross-entropy minimization approach to positron emission tomographic (PET) imaging is described. An unexplored multicriterion cross-entropy optimization algorithm based on weighted-sum scalarization is used to solve this problem. The efficacy of the algorithm is compared with that of the single-criterion optimization algorithm and the convolution backprojection method for image reconstruction from computer-generated projection data and Siemens PET scanner data. The algorithms described have been implemented on a PIII/686 microcomputer.     

Mart Algorithms for Circular and Helical Cone-Beam Tomography

The present work is concerned with the evaluation of the performance and the efficient implementation of multiplicative algebraic reconstruction technique (MART) to reconstruct three-dimensional (3D) objects for two different source/detector trajectories. Three types of MART algorithms are tested on a numerical phantom (Defrise), and they are implemented on a 3D X-ray system of Vikram Sarabhai Space Centre (VSSC). Circular and helical cone-beam trajectories are used. The results are compared with convolution backprojection (CBP) algorithm for each trajectory. It is found that iterative algorithms perform better than their counterpart, the transform-based CBP algorithm, whenever tomography systems are ill-conditioned due to limited views and/or noisy projection data.     

Multicriteria maximum likelihood neural network approach to positron emission tomography

The emerging technology of positron emission image reconstruction is introduced in this paper as a multicriteria optimization problem. We show how selected families of objective functions may be used to reconstruct positron emission images. We develop a novel neural network approach to positron emission imaging problems. We also studied the most frequently used image reconstruction methods, namely, maximum likelihood under the framework of single performance criterion optimization. Finally, we introduced some of the results obtained by various reconstruction algorithms using computer‐generated noisy projection data from a chest phantom and real positron emission tomography (PET) scanner data. Comparison of the reconstructed images indicated that the multicriteria optimization method gave the best in error, smoothness (suppression of noise), gray value resolution, and ghost‐free images. © 2001 John Wiley & Sons, Inc. Int J Imaging Syst Technol 11, 361–364, 2000     

一种新的频谱分析图像重建算法及模拟

这种频谱分析图像重建算法利用二维傅里叶级数的各正交基函数的投影向量,去逼近图像 函数的投影向量,利用优化准则,得到各正交基函数的权系数,再经过傅里叶反变换,即可重建出原始的图像函数。计算机模拟表明,该算法重建时间不足1秒,选择合适的近似级数,对复杂场的图像重建,也能使最大误差不超过5%,平均误差不超过1%。     

基于梯度投影稀疏重建算法的电容层析成像图像重建

为解决两相流中存在中心物体、物体比较小或存在多个物体且相距较近时电容层析成像(ECT)重建图像精度较差的问题,基于稀疏分布的流型其介电常数分布满足稀疏性的先验条件,采用梯度投影稀疏重建(GPSR-BB)算法进行ECT图像重建。仿真及实验测试结果表明:GPSR-BB算法对于流体中小目标以及复杂流型的图像重建质量较好,重建图像的形状保真度高。     

An Efficient SQUID NDE Defect Detection Approach by?Using?an?Adaptive?Finite-Element?Modeling

Incorporating the finite-element method for the modeling of the SQUID NDE response to a predefined defect pattern, an adaptive algorithm has been developed for the reconstruction of unknown defects using an optimization algorithm for updating of the forward problem. The defect reconstruction algorithm starts with an initial estimation for the defect pattern. Then the forward problem is solved and the obtained field pattern is compared with the measured signal from the SQUID NDE system. The result is used by an optimization algorithm to update the defect structure to be incorporated in the FEM forward problem for the next iteration. Since the mentioned model based inverse algorithm normally consumes a lot of computational resources, the number of iterations plays an important role in the determination of the total response convergence time. Consequently, different optimization algorithms have been applied and their performances are compared. In this work by incorporating an efficient forward model and using the stochastic and deterministic optimization algorithms for defect updating we have investigated their performance on the inversion of the SQUID NDE signal and also their ability to defect reconstruction in the noisy environment.     

The integrated acceleration of the Chambolle-Pock algorithm applied to constrained TV minimization in CT image reconstruction

Background and Objective: The constrained, total variation (TV) minimization algorithm has been applied in computed tomography (CT) for more than 10 years to reconstruct images accurately from sparse-view projections. Chambolle-Pock (CP) algorithm framework has been used to derive the algorithm instance for the constrained TV minimization programme. However, the ordinary CP (OCP) algorithm has slower convergence rate and each iteration is also time-consuming. Thus, we investigate the acceleration approaches for achieving fast convergence and high-speed reconstruction. Methods: To achieve fast convergence rate, we propose a new algorithm parameters setting approach for OCP. To achieve high-speed reconstruction in each iteration, we use graphics processing unit (GPU) to speed-up the two time-consuming operations, projection and backprojection. Results: We evaluate and validate our acceleration approaches via two-dimensional (2D) reconstructions of a low-resolution Shepp–Logan phantom from noise-free data and a high-resolution Shepp–Logan phantom from noise-free and noisy data. For the three-specific imaging cases considered here, the convergence process are speeded up for 89, 9 and 81 times, and the reconstruction in each iteration maybe speeded up for 120, 340 and 340 times, respectively. Totally, the whole reconstructions for the three cases are speeded up for about 10,000, 3060 and 27,540 times, respectively. Conclusions: The OCP algorithm maybe tremendously accelerated by use of the improved algorithm parameters setting and use of GPU. The integrated acceleration techniques make the OCP algorithm more practical in the CT reconstruction area.     

Crash reconstruction of pedestrian accidents using optimization techniques

Numerical simulations of vehicle-to-pedestrian crash (VPC) are frequently used to develop a detailed understanding of how pedestrian injuries relate to documented vehicle damage. Given the complexity of the event, modeling the interactions typically involves subjective evaluations of the pre-impact conditions using a limited number of simulations. The goal of this study is to develop a robust methodology for obtaining the pre-impact pedestrian posture and vehicle speed utilizing multi-body simulations and optimization techniques. First, a continuous sequence of the pedestrian gait based on the literature data and simulations was developed for use as a design parameter during the optimization process. Then, the robustness and efficiency of three optimization algorithms were evaluated in a mock (idealized) crash reconstruction. The pre-impact parameters of the pedestrian and the vehicle models were treated as unknown design variables for the purpose of validating the optimization technique. While all algorithms found solutions in close vicinity of the exact solution, a genetic algorithm exhibited the fastest convergence. The response surfaces of the objective function showed higher sensitivities to the pedestrian posture and its relative position with respect to the vehicle than to the vehicle speed for the chosen design space. After validating the methodology with the mock reconstruction, a real-world vehicle-to-pedestrian accident was reconstructed using the data obtained from the field investigation and the optimization methodology. A set of pedestrian and vehicle initial conditions capable of matching all observed contact points was determined. Based on the mock and real-world reconstructions, this study indicates that numerical simulations coupled with optimization algorithms can be used to predict pedestrian and vehicle pre-impact conditions.     

Investigation of smooth basis functions and an approximated projection algorithm for faster tomography

To decrease the calculation time of iterative tomographic reconstruction, this study proposes and investigates the uses of the “dot‐projection” algorithm, which uses smooth basis functions for data representation. Basis functions give an approximation to the basis functions of blobs, which provide a more accurate reconstruction at the same resolution than for traditional “ray‐driven” projection algorithms with voxels. This study evaluates the calculation time decrease and obtained errors from the algorithm for reconstruction, as well as its response to spatial frequencies. Finally, we compare the dot‐projection for several basis functions and the ray‐driven projection with respect to accuracy. © 2001 John Wiley & Sons, Inc. Int J Imaging Syst Technol 11, 347–354, 2000     

超声反射CT在水下物体成像中的应用

本文重点介绍了沿平面扫描成像，提出了修正扫描起始点位置以及换能器孔径影响的一种方法，得到较好的实验结果。