关于分布函数极大值位置的讨论

将关于某个变量的连续分布函数进行变数变换,除非新老变数为线性关系,否则变换前后的分布函数的极大值之间的关系一般会不同于新老变数之间的关系。     

The relation between the generalised Eshelby integral and the generalised BCS and DB modelsvspace1mm

The generalised BCS dislocation group model and the generalised DB atomic cohesive force zone model have obtained the same results on nonlinear fracture study of some one-, two- and three-dimensional quasicrystals. This work reveals some inherent connection between the two models, and finds that their common basis is the generalised Eshelby integral based on the generalised Eshelby energy--momentum tensor for quasicrystals. Further applications of the theory in solving nonlinear fracture problems of the materials are also discussed.     

Modeling of the Temporally and Spatially Modulated Fourier Transform Imaging Spectrometer working in orbit

Temporally and Spatially Modulated Fourier Transform Imaging Spectrometer (TSMFTIS) is a new kind of imaging spectrometer and is applicable in remote sensing. Compared with Temporally Modulated Imaging Interferometer and Spatially Modulated Imaging Interferometer, it employs immovable parts and abandons slit. Based on the analysis of its physical theory, optical principle, data gathering and processing, the mathematical model of the TSMFTIS working in orbit is proposed in this paper. The model is comprised of different function blocks, such as fore-optics, interferometer and so on. For simulation, a big remote sensing image is utilized as ground scene, and a conversion which is similar to color mode conversion from RGB to Lab is made to produce artificial spectra of various ground objects. At last, a validation experiment is performed with real-world spectra of four pieces of colorful cloths by using a commercially available spectrometer.     

基于两步正交相移干涉的振幅图像光学加密技术

提出一种基于两步正交相移干涉的光学图像加密技术.这种相移干涉数字全息只要记录两幅干涉图,不需要记录物光波和参考光波的强度信息,就可以再现没有零级像和共轭像的再现像.物光波对应的光路经过两次菲涅尔变换,并结合双随机相位编码.参考光分别引入0和π/2相位,用数字化记录介质记录两幅数字全息图作为加密图像.解密时只要获得正确的密钥,经过简单的计算就可以重建清晰的原始图像.模拟实验验证了它的可行性和有效性,分析了抗裁剪和噪音的鲁棒性.     

时间分辨傅里叶变换红外光谱法研究聚偏氟乙烯的晶变

本文用时间分辨傅里叶变换红外光谱（ＴＲ－ＦＴＩＲＳ）法研究了聚偏氟乙烯（ＰＶＤＦ）的晶型转变，并认５０８ｃｍ＾－１峰的增强和５２８ｃｍ＾－１峰的减弱表示ａ晶型向β晶型的转变，瞬变的初期时间约５ｍｓ，在晶型转变中伴随有“磁滞回线”的非线性效应。     

基于白光干涉彩色图像的微结构表面形貌测量

微结构的表面形貌会显著地影响微纳器件的使用性能及产品质量,是微纳测试领域的一个重要研究方面,利用白光干涉技术是测量物体表面形貌的一种常见方法。区别于常用的CCD黑白相机,使用CCD彩色相机采集白光干涉条纹的彩色图像,使获取的图像包含了R、G、B三个通道的信息。利用小波变换法分别求解出在不同扫描位置处R、G、B通道的相位信息,通过建立的评价函数,并结合最小二乘法可精确确定零光程差的位置,利用相对高度和零光程差位置的线性关系,进而得到物体的表面形貌。通过仿真以及实际测量由VLSI标准公司制造的标准台阶结构,验证了所提出方法的有效性。     

Full-wave modeling of piezoelectric transducers for SAW acousto-optical interactions

In this paper we present an accurate and efficient numerical method for a rigorous full-wave analysis of interdigital transducers (IDT) for the excitation of surface acoustic waves on the piezoelectric substrate of acousto-optical devices. The problem is formulated in terms of an integral equation that is solved by the method of moments. The transducer input admittance and the power coupling factors to both surface and bulk waves are computed. Numerical results for some configurations of X-Y LiNbO₃ IDT for acousto-optic applications are in very good agreement with measured data. It is pointed out that bulk wave excitation may be a serious limitation in the design of efficient, wide band transducers for acousto-optical devices.     

一种用于衍射光元件优化设计的快速算法的研究

讨论了傅里叶迭代算法对初始相位敏感的缺陷机理,提出用伪随机相位编码作为傅里叶迭代算法初始值的改进算法来克服相位敏感缺陷。改进算法具有加快迭代收敛速度、提高恢复图像强度准确性的优点。将该算法产生的任意期望图样的相位全息图写入纯相位液晶空间光调制器上,在远场成功地再现了期望图样。算法迭代10次可产生衍射效率接近85%的一幅相位全息图。作为一种快速相位恢复算法它在相控阵激光雷达以及光束敏捷偏转等领域有重要意义。     

单层元胞结构金属泡沫的透射特性

本文构建了开孔泡沫铝的简化几何模型,并利用有限积分法模拟了具有单层元胞结构的金属泡沫材料在线性极化平面波垂直入射情况下的透射率。基于菲涅耳-基尔霍夫衍射理论分析了孔隙率、孔径尺寸、材料厚度和骨架结构对金属泡沫材料辐射特性的影响。当尺度参数较大时,金属泡沫材料的固相支架结构满足良导体条件,宏观电磁屏蔽效应显著,金属泡沫材料呈现"非透明"性质。随着入射电磁波波长逐渐接近于孔径尺寸,散射效应越来越显著,金属泡沫材料的"半透明"性质开始显现,不同孔隙率的金属泡沫材料的透射率以相近的规律随波长变化。随着波长的进一步减小,衍射效应对于金属泡沫材料透射特性的影响逐渐占据主导地位,采用菲涅耳-基尔霍夫衍射理论可以较好地描述透射能流在孔隙结构内的分布。当衍射效应占据主导地位时,对于相同孔隙率金属泡沫材料,孔径尺寸对衍射光学行为影响不大,而材料厚度、孔隙率和骨架结构会显著影响金属泡沫材料的透射率。     

基于二代curvelet变换的图像融合研究

曲波(Curvelet)作为一种新的多尺度分析方法比小波更加适合分析二维图像中的曲线或直线状边缘特征,而且具有更高的逼近精度和更好的稀疏表达能力.将curvelet变换引入图像融合,能够更好地提取原始图像的特征,为融合图像提供更多的信息.第二代curvelet理论的提出也使得其理论更易理解和实现.因此,提出了一种基于第二代curvelet变换的图像融合方法,首先将图像进行curvelet变换,然后在相应尺度上利用融合规则将curvelet系数融合,最后进行重构得到融合结果.对多聚焦图像进行了实验,采用均方误差、偏差指数和相关系数对融合结果进行了客观评价,并与基于小波变换的融合进行了比较,实验结果表明该方法除分解2层时与小波性能相当,取其他分解层数时均获得更好的融合效果.