基于投影融合的多层球壳双能CT算法研究

对高密度差多层球壳结构件进行双能CT检测，需要研究同时基于高低能射线两套投影数据的图像重建方法。先分别重建再进行图像融合的方法，不能充分利用两种能量射线投影的互补信息，重建图像质量不高。针对高密度差多层球壳结构特征，通过设计限定投影扫描方式获取区域分布明确的高、低能射线投影正弦图，基于正弦图区域分割重组和交集数据等值化一致性处理实现双能投影数据融合，在此基础上完成图像重建。根据以上设计归纳出基于投影融合的多层球壳双能CT重建算法，仿真结果表明算法可行，且重建结果明显好于基于图像融合的方法。     

基于互补色编码条纹投影的三维形貌测量方法

针对彩色编码条纹投影轮廓术中,被测物体表面颜色易与投影条纹颜色发生干扰的问题,提出了一种新的基于互补色编码条纹投影的三维测量方法.先后向被测物体投影一幅主彩色编码正弦条纹图和一幅与前者包含的彩色编码条纹颜色互为补色的彩色条纹图.利用两者颜色互补的特点,判断两幅图像中对应像素点的颜色状态来确定其在主彩色编码正弦条纹图中所...     

基于投影匹配的X射线双能计算机层析成像投影分解算法

X射线双能计算机层析成像(CT)技术是安全检查领域一种重要的材料探测与识别手段.双能CT投影分解是双能CT预处理重建算法的核心内容和关键步骤.针对现有投影分解算法的不足,提出了一种基于投影匹配的双能CT投影分解算法.依据系统能谱和基材料线性衰减系数曲线,通过求解投影积分方程组建立高低能投影查找表.对于给定的高低能投影,...     

基于点阵投影的彩色复合光栅傅里叶变换轮廓术

传统傅里叶变换轮廓术(FTP)采用单一频率条纹测量时存在着由于待测物体高度变化剧烈导致的相位展开困难。结合数字点阵和正弦条纹投影的复合编码三维数字成像方法能有效解决相位模糊和误差传播的问题,但在实际测量中需要分别采集点阵图像和正弦条纹图像,影响了测量的实时性。提出一种新型彩色复合光栅测量方法,实现了从一帧图像中分别得到点阵图像和正弦条纹图像,并以点阵图像为依据对正弦条纹图的截断相位进行展开。理论分析和实验结果表明,该方法适合于不连续物体的三维测量,同时抑制误差传播能力强,具有较高的测量效率。     

基于机器学习的束团横向相空间测量

理论上，使用断层扫描技术可以得到真实的横向相空间分布。但是想要更加精确地了解分布的细节，需要解决旋转角度范围受限和投影数目不足的问题。针对这两个问题，提出了在混合域处理的神经网络模型，即组合地在正弦域和断层域分别使用插值和去除伪影神经网络。在简单地测量束线以及投影数目比较少（7个）的情况下，该网络模型也能高质量地重建束团横向相空间分布。并且，由于选择旋转角度的方式和归一化相空间无关，因此，无需测量Twiss参数。采用该方法测量束团横向相空间，一定程度提升了重建质量，简化了测量的方式。     

点云投影结合轻量化卷积神经网络实现三维成像声呐快速目标分类*

三维成像声呐的成像结果是三维点云,基于点云的三维成像声呐目标分类方法具有网络结构复杂,计算量大的特点,针对这一问题本文提出了一种将三维成像声呐成像结果从三维点云投影至二维图像的方法,并且使用轻量化卷积神经网络实现了三维成像声呐快速目标分类。该方法首先对三维成像声呐波束形成后的波束域数据进行最大值滤波和阈值滤波,降低点云数据维度;接着,依据三维成像声呐的波束方向,将点云投影为深度图和强度图,分别保存点云的位置信息和强度信息;然后,利用深度图和强度图分别作为第一个通道和第二个通道构建混合通道图,将混合通道图作为目标分类网络的输入,从而将三维点云的目标分类问题转换为二维图像的目标分类问题;最后使用MobileNetV2网络实现了三维成像声呐快速目标分类。实验结果表明,通过本文提出的投影方法可以用二维图像分类网络完成三维成像声呐点云的目标分类任务;而且混合通道图比单独的强度图和深度图收敛速度更快,结合目标识别网络可以25fps实时的进行目标分类,在真实数据集上分类精度达到了91.13%。     

航空相机动态调制传递函数分析与研究

利用调制传递函数对动态成像质量做出评价,确立合理的稳定系统性能指标,是航空相机设计过程中需要解决的重要问题.基于线性光学系统的传递函数理论,针对匀速运动、加速运动和正弦运动等几种动态成像条件进行了定量分析,建立了像移量和动态成像质量之间的联系.针对低频正弦运动,提出了一种基于离散相位区间的动态调制传递函数计算方法.分析结果表明,在像模糊斑大小均为0.5个像元的情况下,高频正弦运动、直线运动和低频正弦运动的动态调制传递函数分别可以达到0.85、0.9和0.998,此时可以获得较理想的动态图像.据此得出了高频振动对像质的破坏最严重,直线运动次之,低频正弦运动最轻的结论.根据分析,可以针对特定的动态成像条件计算相应的调制传递函数,为航空相机系统的综合像质评价和视轴稳定系统设计提供了准确的定量参考.     

真实场景的三维视频采集及显示

提出了一种真实场景三维视频采集及彩色显示的方法.设计了一种采用条纹投影的实时三维成像系统及采用液晶空间光调制器的实时全息彩色三维显示系统.在三维成像系统中采用π相移正弦条纹与编码图案结合实现绝对相位测量,从而可以测量孤立物体.同时对采用数字微镜的投影仪进行改造,实现高速投影,并与高速摄像机配合实现三维视频采集.首先利用...     

基于傅里叶频谱分析的相位测量轮廓术系统Gamma非线性校正方法

数字投影和成像系统的Gamma非线性效应是导致相位测量轮廓术（PMP）测量误差的重要原因。目前大多采用多帧条纹图进行Gamma校正,使测量的实时性受到限制。提出了一种Gamma校正方法,利用正交光栅像的傅里叶频谱分布计算Gamma值,再根据此Gamma值对投影相移条纹进行Gamma逆变换,实现投影条纹输入值的提前校正,以获取具有良好正弦性的结构条纹,从而降低PMP相位测量误差。校正过程中只需一帧条纹图,而且考虑了测量系统的离焦效应。实验证实了该方法的有效性和正确性。     

单像素探测成像实验研究

提出了不需要使用任何几何光学成像系统的单像素探测成像方法并进行实验研究.通过投影一系列不同频率组合、且每种频率组合具有4种不同初相位的二维正弦灰度条纹到成像物体表面,利用单像素探测器探测来自成像物体的后向散射光,然后对成像物体的频谱进行重建.本文给出详细的理论基础、实验装置和实验结果.     

Generalized Titarenko's algorithm for ring artefacts reduction

A fast algorithm for ring artefact reduction in high‐resolution micro‐tomography with synchrotron radiation is presented. The new method is a generalization of the one proposed by Titarenko and collaborators, with a complete sinogram restoration prior to reconstruction with classical algorithms. The generalized algorithm can be performed in linear time and is easy to implement. Compared with the original approach, with an explicit solution, this approach is fast through the use of the conjugate gradient method. Also, low/high‐resolution sinograms can be restored using higher/lower‐order derivatives of the projections. Using different order for the derivative is an advantage over the classical Titarenko's approach. Several numerical results using the proposed method are provided, supporting our claims.     

一种基于改进频域图像配准技术的超分辨力图像处理方法

为提高光电成像系统的空间分辨力,提出了一种基于改进的频率域图像配准技术的超分辨力图像处理方法。首先利用改进的频域图像配准方法估算出低分辨力图像之间的微位移量,然后采用Papoulis-Gerchberg超分辨力处理方法完成图像复原。利用不同重构方法进行了仿真及实验研究,给出了评价参数。模拟和实际显微热图像的处理结果表明:该算法可使图像质量得到改善,分辨的细节更多,可有效地提高光电成像系统的空间分辨力;处理算法简单,计算量小,可实现快速处理。该算法还可应用于其他不可控光学微扫描成像系统中,具有广泛的应用前景。     

Deep learning for image reconstruction in thermoacoustic tomography

Microwave-induced thermoacoustic tomography(TAT)is a rapidly-developing noninvasive imaging technique that integrates the advantages of microwave imaging and ultrasound imaging.While an image reconstruction algorithm is critical for the TAT,current reconstruction methods often creates significant artifacts and are computationally costly.In this work,we propose a deep learning-based end-to-end image reconstruction method to achieve the direct reconstruction from the sinogram data to the initial pressure density image.We design a new network architecture TAT-Net to transfer the sinogram domain to the image domain with high accuracy.For the scenarios where realistic training data are scarce or unavailable,we use the finite element method(FEM)to generate synthetic data where the domain gap between the synthetic and realistic data is resolved through the signal processing method.The TAT-Net trained with synthetic data is evaluated through both simulations and phantom experiments and achieves competitive performance in artifact removal and robustness.Compared with other state-of-the-art reconstruction methods,the TAT-Net method can reduce the root mean square error to 0.0143,and increase the structure similarity and peak signal-to-noise ratio to 0.988 and 38.64,respectively.The results obtained indicate that the TAT-Net has great potential applications in improving image reconstruction quality and fast quantitative reconstruction.     

A fast digital image correlation method for deformation measurement

Fast and high-accuracy deformation analysis using digital image correlation (DIC) has been increasingly important and highly demanded in recent years. In literature, the DIC method using the Newton-Rapshon (NR) algorithm has been considered as a gold standard for accurate sub-pixel displacement tracking, as it is insensitive to the relative deformation and rotation of the target subset and thus provides highest sub-pixel registration accuracy and widest applicability. A significant drawback of conventional NR-algorithm-based DIC method, however, is its extremely huge computational expense. In this paper, a fast DIC method is proposed deformation measurement by effectively eliminating the repeating redundant calculations involved in the conventional NR-algorithm-based DIC method. Specifically, a reliability-guided displacement scanning strategy is employed to avoid time-consuming integer-pixel displacement searching for each calculation point, and a pre-computed global interpolation coefficient look-up table is utilized to entirely eliminate repetitive interpolation calculation at sub-pixel locations. With these two approaches, the proposed fast DIC method substantially increases the calculation efficiency of the traditional NR-algorithm-based DIC method. The performance of proposed fast DIC method is carefully tested on real experimental images using various calculation parameters. Results reveal that the computational speed of the present fast DIC is about 120-200 times faster than that of the traditional method, without any loss of its measurement accuracy     

利用先验估计自适应加窗的准静态超声弹性成像位移估计算法

准静态弹性成像技术是基于组织压缩前和压缩后超声回波射频信号进行组织运动重构的弹性成像技术。提出了一种基于先验估计的自适应窗函数算法,在位移估计过程中,使用已估计的临近窗的时延值作为先验信息,自动调整截取压缩后射频信号段的截取窗函数,提高了互相关运算所需的压缩前和压缩后信号段之间的相关性。仿真实验结果表明,该算法不仅大大提高了成像速度,而且提高了信噪比较低时的成像质量,同时该算法具有更宽的应变通带。      

低信噪比下的二维联合线性布雷格曼迭代快速超分辨成像算法

针对实际逆合成孔径雷达(ISAR)成像时带宽有限、方位孔径稀疏的小角度回波数据条件下,常规算法的成像分辨率不高等问题,基于压缩感知理论,提出了一种低信噪比条件下的二维联合布雷格曼迭代快速ISAR超分辨成像算法.首先,将雷达回波构建为距离频域-方位多普勒域的二维稀疏表示模型,在此基础上,将二维超分辨成像问题转换为二维联合压缩感知的稀疏重构问题;其次,为了避免重构时向量化操作带来的复杂度,提出了二维联合布雷格曼迭代算法,为实现快速重构,将加权残量迭代、估计停滞步长与感知矩阵条件数优化三种加快收敛速度的思想相结合,既利用了布雷格曼迭代在低信噪比条件下的重构能力又能保证快速成像.最后仿真实验结果表明在欠采样和低信噪比条件下本文算法能够缩短成像时间,且具备更好的噪声鲁棒性.     

Random noise filtering method based on the inter-frame registration

One of the problems to be solved in image processing is how to eliminate image noise effectively. In this work, we brought forward a random noise filtering method based on the inter-frame registration. Firstly, we calculated the relative displacement of the adjacent frames by a registration algorithm. Then we divided the image into the overlapping area and the non-overlapping area according to the relative displacement. Finally, we do noise reduction processing for these two areas respectively. The experiments results indicate that the proposed method can reduce noise in both spatial and time domain of video images. The main advantage is that it cannot only remove noise, but also effectively protect the image edge and detail information. Besides, it not only maintains the de-noising effect of traditional inter-frame algorithm, but also is suitable for moving targets. It has better real-time performance and wider application range.     

Digital speckle correlation method with compensation technique for strain field measurements

In the strain field measurements, especially in large strain field measurements, the correlation coefficient values obtained by the digital speckle correlation method are usually very low due to the relative pixel movement of the subset. The measuring error, therefore, is increased. A primary method, that is called compensation algorithm, is introduced for improving the correlation coefficient. A flow scheme with the compensation algorithm of our software is developed and some improved techniques for reducing the calculation time and error are discussed in the paper. After obtaining a set of displacement data with high correlation coefficient, a de-noise wavelet processing is adopted. It is obvious that the measuring accuracy of the strain field is better then before. A strain field of testing experiment is performed with this compensation technique. The correlation coefficients can increase from 0.70 to 0.99, which will be of much benefit for the improvements of the measuring accuracy.     

Tissue velocity imaging of coronary artery by rotating-type intravascular ultrasound

Intravascular ultrasound (IVUS) provides not only the dimensions of coronary artery but the information of tissue components. In catheterization laboratory, soft and hard plaques are classified by visual inspection of echo intensity. So-called soft plaque contains lipid core or thrombus and it is believed to be more vulnerable than a hard plaque. However, it is not simple to analyze the echo signals quantitatively. When we look at a reflection signal, the intensity is affected by the distance of the object, the medium between transducer and objects and the fluctuation caused by rotation of IVUS probe. The time of flight is also affected by the sound speed of the medium and Doppler shift caused by tissue motion but usually those can be neglected. Thus, the analysis of RF signal in time domain can be more quantitative than intensity of RF signal. In the present study, a novel imaging technique called "intravascular tissue velocity imaging" was developed for searching a vulnerable plaque. Radio-frequency (RF) signal from a clinically used IVUS apparatus was digitized at 500 MSa/s and stored in a workstation. First, non-uniform rotation was corrected by maximizing the correlation coefficient of circumferential RF signal distribution in two consecutive frames. Then, the correlation and displacement were calculated by analyzing the radial difference of RF signal. Tissue velocity was determined by the displacement and the frame rate. The correlation image of normal and atherosclerotic coronary arteries clearly showed the internal and external borders of arterial wall. Soft plaque with low echo area in the intima showed high velocity while the calcified lesion showed the very low tissue velocity. This technique provides important information on tissue character of coronary artery.     

Image improvement based on sub-pixel post-integration for a staring imaging system in geostationary orbit

This paper proposes the post-integration technology based on sub-pixel image registration and image fusion to improve the signal-to-noise ratio （SNR） of remote sensing images without motion degradation caused by satellite vibration. A two-dimensional vibration system is set up to simulate satellite disturbance. Image sequences with different exposure times are captured using a high-speed CMOS camera. The displacement plots are compared with the motion data measured by the grating linear encoder. These plots indicate that the accuracy of the registration algorithm is better than 0.1 pixels. The sub-pixel image fusion shows an improvement in image quality, thus indicating that this technology is powerful for staring imaging systems in geostationary orbit.