反射式鬼成像在海洋湍流中的成像分析

基于惠更斯-菲涅尔原理和鬼成像理论,研究了反射式无透镜鬼成像在Kolmogorov海洋湍流中目标成像问题,得到了成像脉冲响应函数和可见度理论表达式。结果表明随着目标随机反射角的增大,关联重构的图像质量将会失真退化.为进一步分析海洋湍流对目标图像质量的影响,数值模拟了不同探测距离、不同海洋湍流情况下鬼成像的可见度,得到了海洋湍流作用下的鬼成像目标探测规律.该研究对远距离自适应水下鬼成像大规模应用具有重要的理论指导意义.     

反射式光纤共焦扫描成像的研究

建立了反射式光纤共焦扫描成像系统,分析了光纤－集光透镜参数Ａ及物透镜有效数值孔径等对系统成像分辨率的影响。并在此基础上选择了合适参数的透镜,获得了优化的反射式光纤共焦扫描成像系统,测试结果表明,该系统具有亚微米级横向成像能力,微米级向层析能力,成像稳定性那,它将应用于材料及生物组织三维成像检测中。     

基于级联卷积神经网络的前列腺磁共振图像分类

针对深度学习训练成本高,以及基于磁共振图像的前列腺癌临床诊断需要大量医学常识且极为耗时的问题,本文提出了一种基于级联卷积神经网络（Convolutional Neural Network,CNN）和磁共振图像的前列腺癌（Prostate Cancer,PCa）自动分类诊断方法,该网络以Faster-RCNN作为前网络,对前列腺区域进行提取分割,用于排除前列腺附近组织器官的干扰;以基于ResNet改进的网络结构CNN40_bottleneck作为后网络,用于对前列腺区域病变进行分类.后网络由瓶颈结构串联组成,其中使用批量标准化（Batch Normalization,BN）、全局平均池化（Global Average Pooling,GAP）进行优化.实验结果证明,本文方法对前列腺癌诊断结果较好,而且缩减了训练时间和参数量,有效降低了训练成本.     

基于深度递归级联卷积神经网络的并行磁共振成像方法

快速磁共振成像是磁共振研究领域重要的课题之一.随着大数据和深度学习的兴起,神经网络成为快速磁共振技术的重要方法.然而网络性能表现和网络参数量之间较难取得平衡,且对于多通道数据重建的并行成像问题,相关研究较少.本文构建了一种深度递归级联卷积神经网络结构,用于处理并行成像问题.这种网络结构在减少网络参数量的同时,能够尽可能地提高网络的表达能力,提高网络重建的精确度.实验结果表明,相较于传统并行成像方法,通过训练好的神经网络对欠采样磁共振数据进行重建,可以得到更准确的重建结果,且重建时间大大缩短.     

基于鬼成像的手写字体分类方法

针对手写字体特征提取难、识别效率低甚至错误的问题,提出一种结合鬼成像原理和分类网络的思想用于手写字体的识别。首先,使用激光器、数字微镜阵列、单像素探测器等搭建鬼成像探测系统,利用系列哈达玛散斑场分别实现对17 239幅手写字体的照射;其次,利用单像素探测器对手写字体透射的总光强值进行采集,并将其转换为一维向量作为网络训练的输入;最后基于卷积神经网络在图像分类的优势搭建网络架构且为解决训练过程中的网络退化问题加入残差块。实验结果表明：对于手写数字来说,全连接神经网络与卷积神经网络的精密度、召回率、F1值分别提高86.50%/97.25%、86.40%/98.03%、86.31%/97.60%;对于手写字母来说,卷积神经网络在全采样下的精密度、召回率、F1值分别为91.87%、90%、90.23%。该方法利用鬼成像原理,仅通过手写字体透射的总光强值而无需对手写字体的特征进行提取与识别便可进行快速分类,大幅提升了手写字体的识别效率。     

结合高光谱成像和机器学习的棉种年份鉴别

棉花精量播种技术目前已经在新疆兵团全面推广,该技术能精确实现一穴一粒的农艺技术指标,但是也对高质量棉种的筛选提出了更高的要求。为了避免播种往年活力不足的棉种而导致发芽率降低的问题,结合机器学习和近红外（NIR）高光谱成像技术（HSI）进行棉种年份精确鉴别,实现棉种的快速无损筛选。采集2016年—2019年近四年外观无明显差异的棉种各360粒,共1 440粒棉种（按照3∶1∶1划分训练集、验证集和测试集）作为样本,按照每批60粒采集915~1 698 nm范围的棉种高光谱图像,去除首尾两端噪声大的光谱,保留1 002~1 602 nm范围的光谱为原始数据。利用Savitzky-Golay（SG）平滑算法对光谱进行预处理,采用主成分载荷方法（PCA-loading）选取13个特征波段,基于全部光谱数据和特征波段（±10 nm）数据建立逻辑回归（LR）、偏最小二乘判别分析（PLS-DA）、支持向量机（SVM）、循环神经网络（RNN）、长短记忆网络（LSTM）和卷积神经网络（CNN）六种分类模型。使用全光谱数据建模时,六种分类模型在测试集上的鉴别准确率分别为96.27%,98.98%,99.32%,96.95%,97.63%和100%,其中CNN和SVM模型取得了较好的结果;使用特征光谱数据建模时,六种分类模型在测试集上的鉴别精度分别为93.56%,97.29%,98.30%,95.25%,94.24%和99.66%,其中CNN和SVM模型仍有较好的分类结果。结果表明,使用全光谱数据建模时,六种分类模型都可以实现较高精度的棉种年份鉴别,使用特征光谱数据建模时CNN和SVM模型的鉴别精度仍可达到98%;其中深度学习方法优于传统机器学习方法,但是传统机器学习方法仍能保持较好的鉴别准确率。因此,结合近红外高光谱成像技术和机器学习方法能够实现棉种年份的高精度鉴别,为棉花精量播种过程中的优质棉种选种技术提供理论依据和方法。     

核燃料棒反光表面条纹自适应中心提取方法

针对水下检测中激光条纹中心难以准确提取的问题,提出了一种适用于水下核燃料棒反光表面的条纹自适应中心提取方法.根据检测环境中存在水体散射及物体表面高反光的特点,去除水下噪点、反光噪点,实现激光条纹分割提取;充分利用BP神经网络曲线拟合及由光条几何信息生成的自适应卷积模板,实现反光区域的轮廓与灰度分布修正,使光条截面灰度分...     

超强边缘磁场梯度场下的磁共振成像:STRAFI 实验

磁共振成像（MRI）是一个能够探测样品内部特性的有效检测手段,已被广泛应用于化学、生物研究,以及医疗诊断领域. 自约40年前发展以来, 成像方法的不断发展使得MRI的成像分辨率、实验效率和成像杂核能力得到了很大的改进. 边缘磁场成像（STRAFI）是一种很具潜力的成像方法之一,它利用了超导磁体本身具有的边缘场的强梯度场. 该综述介绍了STRAFI基础,并概括了成像的基本原理、STRAFI的实验理论和方法及其在实际研究中的应用. 由此将比较STRAFI实验相对于传统MRI方法的所具有的优势和多面可行性.     

高空间分辨率高可见度的太赫兹光谱成像研究

太赫兹光谱成像,不但包括在二维图像空间的强度信息,同时可以得到太赫兹波段的光谱信息,构成了一个三维的数据矩阵。由于受到太赫兹成像系统内部硬件的限制和影响,太赫兹频域较高频段处信号存在能量弱、信噪比低的特点,导致所成的太赫兹图像普遍存在分辨率低、对比度低等问题。因此,利用三维数据矩阵,应用适合的算法,实现了提高太赫兹光谱成像空间分辨率、边缘细节可见度的目的。搭建了三维可移动式太赫兹时域光谱成像系统,实现了对标准高分辨率板的二维扫描。对该系统所采集到的信号分别进行时域、频域等多种方式成像对比,结合瑞利判据和分辨率标尺对成像系统的空间分辨率、景深进行标定,研究了提高太赫兹光谱成像的空间分辨率算法。然后,针对太赫兹频域高频区域信噪比低、对比度低、噪声原因复杂的特点,结合深度残差学习的图像去噪理论,提出了太赫兹图像深度去噪网络,在训练集中引入成像系统中真实的“太赫兹残差噪声”。最后,利用所训练出的模型对太赫兹频域高频区域图像进行盲去噪,并用重建图像分别与原始成像结果和传统太赫兹去噪算法结果进行比较,分别从主观和客观两个方面评价了不同算法对太赫兹频域高频图像的去噪效果。实验结果表明,通过该算法实现...     

Reflection-type integral imaging scheme for displaying three-dimensional images

A reflection-type integral imaging scheme for displaying three-dimensional images is proposed. By use of a concave mirror array instead of a lens array, three-dimensional images are integrated in the form of a reflection type, and the experimental results are demonstrated. This scheme can readily be applied to a large integral imaging system by use of a beam projector that is located at a distance from the mirror-array plane.     

Experimental method for the extraction of intensity profiles by imaging the scattered pattern with a charge-coupled devic

A practicable experimental method for measuring scattering on rough surfaces is reported.The scattered patterns are captured on a screen composed of two pieces of ground glass and then imaged using a chargecoupled device.The scattered intensity profiles are extracted by converting the patterns in real space into the wave vector space.Isotropic and anisotropic samples of the rough backsides of silicon wafer are investigated respectively,and their intensity profiles are measured.The profiles of the anisotropic sample are obtained by reading the pixels on the specific orientation curves.The parameters of the samples are extracted using angle-resolved light-scattering schemes and theories.The results well agree with measurements obtained using an atomic force microscope.     

Analysis and Improvement of Reflection-type Transverse Modulation Optical Voltage Sensors

Reflection-type transverse modulation optical voltage sensors, which employ reflection retarders to replace quarter-wave plates, are convenient for practical use. In previous literatures, the measured voltage was all applied to Bi4Ge3O12 crystal along the (110) direction for transverse modulation optical voltage sensor, and crystals are used as sensing materials. In this paper, reflection-type transverse modulation optical voltage sensor has been analyzed theoretically and a novel configuration in which the measured voltage is applied to a Bi4C-e3O12 crystal along the (001) direction with light wave passing through the crystal in the (110) direction has been proposed. According to this theoretical analysis, a novel optical voltage sensor, which can be used in a 220 kV optical fiber voltage transformer, has been designed and assembled. Experimental results showed that the linearity and the stability of the sensor during 24 hours can reach 0.3％.     

A multi-resolution wavelet algorithm for hand vein pattern recognition

A novel hand vein recognition algorithm is developed based on multi-resolution wavelet analysis. The texture feature of hand vein can be extracted by three-level wavelet decomposition. Furthermore, Knearest neighbor (KNN) with support vector machines (SVM) and minimum distance classifier (MDC) are employed for feature matching. Finally, the experiments are respectively performed in identification and verification modes using Tianjin University (TJU) hand vein image database constructed by our group.The results show the feasibility and effectiveness of the proposed method.     

Analysis and Improvement of Reflection-type Transverse Modulation Optical Voltage Sensors

1　Ｉｎｔｒｏｄｕｃｔｉｏｎ　　Ｏｐｔｉｃａｌｆｉｂｅｒｖｏｌｔａｇｅｔｒａｎｓｆｏｒｍｅｒｓａｒｅｂｅｉｎｇｄｅｖｅｌｏｐｅｄｆｏｒｕｓｅｉｎｔｈｅｅｌｅｃｔｒｉｃｐｏｗｅｒｓｙｓｔｅｍ[1 ,2 ] ,ｂｅｃａｕｓｅｏｆｔｈｅｉｒｏｕｔｓｔａｎｄｉｎｇａｄｖａｎｔａｇｅｓｓｕｃｈａｓｓｉｍｐｌｅｉｎｓｕｌａｎｔｉｏｎｓｔｒｕｃｔｕｒｅ,ｉｍｍｕｎｉｔｙｔｏｅｌｅｃｔｒｏｍａｇｎｅｔｉｃｉｎｔｅｒｆｅｒｅｎｃｅ ,ｎｏｓａｔｕｒａｔｉｏｎｅｆｆｅｃｔ,ｎｏｆｌａｍｍａｂｌｅｍａｔｅｒｉａｌｓｓｕｃｈａｓｏｉ…     

Near infrared imaging of micro-structured polymer-metal surface pattern

Two-dimensional infrared scanning microscopy images of micro-structured surface patterns in an organic transistor device with metal electrode stripes on a polymer channel layer have been demonstrated. A compact single mode optical fiber scanning probe which has a micro dome-shape lens at the tip end was used for the scanning in which reflected beam intensities are coupled back to the optical fiber transceiver depending on the returned power of the sample material. Based on the experiment, the obtained structural dimensions of the micro structure specimen were well matched to the designed ones. These images were compared to the digital microscopy photos for a potential usage to conduct not only the in situ microscopic electrode pattern monitoring of the device but also to non-destructively investigate any surface reflecting material in micrometer scales.     

Coupling of crystalline colloidal arrays with intrinsically conductive polymers: Reflection-type electrochromic devices

The attenuation of the reflected light from a polystyrene (PS) polymerized crystalline colloidal array (PCCA) is achieved through the electrochemical switching of the electrochromic material polyaniline-poly(2-acrylamido-2-methyl-1-propanesulfonicacid) (PANI-PAMPS). For this purpose, layered electrochromic devices (ECDs) were fabricated that employed a PS PCCA film, acting as both a reflecting substrate and polymeric electrolyte, and the conducting polymer PANI-PAMPS, which served as the electrochromic layer. The resulting ECDs exhibited the attenuation of reflectance peak at the stop band position of the PCCA with a 3% of reflectance contrast ratio.     

Cryolitic vein XPS imaging from an industrial grade carbon

Aluminium carbide formation and subsequent dissolution is a generally accepted mechanism to explain the cathode wear in the Hall-Héroult industry. Carbide formation is thought to occur inside the cathode pores. Being dependent upon cathodic current density, it is believed to be an electrochemical process. It is also associated with the presence of cryolitic bath species involved in the dissolution of the carbide layer and thus enabling further carbide formation.