基于斑点噪声的拖尾Rayleigh分布的合成孔径雷达图像最大后验概率降噪

为了反映合成孔径雷达图像中斑点噪声尖峰厚尾的统计特征，使用拖尾Rayleigh分布来描述斑点噪声.基于Gamma先验分布和斑点噪声的拖尾Rayleigh分布，推导出了合成孔径雷达图像的最大后验概率滤波方程，并给出了它在特定特征参数时的解析形式.使用Mellin变换从观察图像估计拖尾Rayleigh分布的未知参数.给出了在斑点噪声的拖尾Rayleigh分布下的最大后验概率降噪试验和量化指标.为了消除滑动窗大小和噪声强度对降噪结果的影响，给出了降噪能力随滑动窗大小和噪声方差的动态变化关系.结果表明，拖尾Ray 关键词： 斑点噪声 拖尾Rayleigh分布 最大后验概率降噪 Mellin变换     

SAR影像极化特征的混合高斯模型与分类

针对高分辨率极化合成孔径雷达(SAR)影像中极化特征呈现尖峰拖尾等复杂多样的统计特点,采用混合高斯模型(GMM)对极化特征进行建模,提出了一种约束距离的混合多元高斯分布的参数估计算法。该参数估计算法在贪婪期望最大算法框架下设计约束距离函数,自动估计混合分量的个数和模型参数,进而在贝叶斯框架下实现极化SAR影像的地物分类。对Radarsat-2旧金山等地区三组影像数据的分类结果表明:相比于经典的分类算法,所提GMM分类算法的总体精度提高了7%～10%,且对样本数目的依赖性更小,在城区和耕地区域等异质区域可以得到精度更高的分类结果。     

星空观测图像目标拖尾的自动消除

为了实现对星空观测图像中高亮目标产生的拖尾现象进行自动消除,建立了目标拖尾自动消除系统,对该系统所采用的图像背景高斯噪声分布参数估计、目标拖尾检测以及图像灰度值校正等算法进行研究。利用直方图最小二乘曲线拟合估计星空观测图像背景高斯噪声的分布参数;利用观测图像中目标拖尾现象的特征提出一种基于统计信息的拖尾检测算法;在确定目标拖尾位置的基础上对被污染的像素进行灰度值校正;利用Visual Studio 2005开发出一套星空观测图像目标拖尾自动消除系统。实验结果表明:针对16 bit,1 0241 024的星空观测图像,单帧图像处理时间约为300 ms,拖尾现象得到消除,恒星和目标等有用信息未被破坏。本文算法基本满足后续观测图像中弱小目标检测稳定、可靠、精度高等要求。     

海洋环境噪声中的α稳定分布模型

针对海洋环境噪声的非高斯特性,将α稳定分布模型应用到噪声统计建模中。基于模型特征指数与样本峰度物理意义的相似性,建立了二者的对应关系,并推导得出相互独立的两个随机变量之和的峰度。通过数值仿真,表明了α稳定分布对尖峰态噪声建模的可行性和适应性。以安静环境、气枪干扰、航船干扰3种典型环境下的海上实测数据为样本,借助分位数图判断了样本的非高斯特性,并利用分位数法和特征函数法联合估计了模型的参数,得出了特征指数的统计规律:安静环境下特征指数接近2.0,趋近于正态分布;气枪干扰下特征指数介于1.2~1.7之间,具有尖峰厚尾特性;航船干扰下与航速、干扰源到接收位置距离有关,航船影响越显著,特征指数越小,该现象恰好与干扰噪声的峰度变化在物理意义上一致。模型验证结果表明,α稳定分布对3类样本均表现出良好的建模效果,尤其对具有尖峰特性的气枪干扰噪声,效果远远优于正态分布。      

含水体的合成孔径雷达图像配准

水体是合成孔径雷达(SAR)图像解译的一类重要内容。针对含水体的SAR图像的成像特点,给出了一种基于轮廓的配准方法。首先,提出了融合观测图像局部统计信息的自适应权马尔科夫随机场(MRF)分割模型,以分割SAR图像水体目标并提取其精确轮廓。然后,提出了轮廓匹配的非均匀高斯混合模型(GMM),该模型能融合轮廓上点的位置信息和以轮廓点为中心的窗口的灰度相似性信息。最后,对含水体目标的SAR图像进行配准实验。结果显示所提出的MRF分割模型能精确地定位目标边缘并保持图像的细节,轮廓匹配的非均匀GMM对噪声、外点及局部变形具有稳健性,能较好地实现含水体目标的SAR图像配准。     

帧转移型面阵CCD相机拖尾评价标准和校正方法

为了有效校正装配帧转移型面阵CCD的星载偏振相机成像时产生的拖尾,开展拖尾程度评价标准和校正方法的研究.基于装配转移型面阵CCD的多角度偏振成像仪工作原理,分析了拖尾的产生机制和暗行法校正模型的特点,建立了多角度偏振成像仪拍摄积分球的拖尾仿真模型.结合仿真模型提出用拖尾区域的灰度标准差和平均梯度作为图像拖尾程度的评价标...     

基于NSCT变换的高分三号SAR与光学图像融合

高分三号卫星是世界上成像模式最多的合成孔径雷达(SAR)卫星,高分三号SAR图像与多光谱图像融合可以改善图像视觉效果。因此提出一种新的研究思路,即利用非下采样轮廓波变换(NSCT)模拟出既包含多光谱谱段信息又体现SAR图像细节信息的高分辨率图像,则融合可不拘泥于具体算法。同时提出两种基于NSCT的高分辨率图像模拟方法,利用高分三号3 m、5 m分辨率SAR图像和高分一号16 m分辨率图像进行实验,采用不同融合算法验证了该思路的有效性。研究结果表明:传统的SAR和多光谱图像直接融合的方法能够保持SAR的细节信息,但噪声明显,且光谱信息损失大;而所提出的NSCT平均图像和平均NSCT图像可以保留融合结果的光谱信息,且模拟的光谱信息前者比后者更贴近多光谱。     

高分辨率合成孔径雷达图像的Gamma分布下最大后验概率降斑算法

针对传统的Gamma分布下最大后验概率降斑算法不能有效保留均匀区域的点目标, 不能有效保留弱边缘以及不能有效滤除强边缘区域的斑点等问题, 提出了基于第二类统计量的先验参数估计的高分辨率合成孔径雷达图像Gamma 分布下最大后验概率降斑算法. 使用Mellin卷积和斑点的乘性模型, Gamma先验分布的参数可由观察图像的前两阶对数累积量精确估计.所提算法具有解析的滤波输出, 便于实现.农田和城区的高分辨率合成孔径雷达图像的降斑实验表明, 与传统的Gamma分布下最大后验概率降斑算法相比, 所提算法既能有效保留均匀区域的点目标, 又能有效保留弱边缘, 还能有效滤除强边缘区域的斑点. 关键词： 高分辨率合成孔径雷达图像 Gamma分布下最大后验概率降斑算法 第二类统计量 对数累积量     

适用于推扫式成像光谱仪的CCD拖尾扣除方法

拖尾(smear)现象是面阵CCD器件的固有问题。在推扫式成像光谱仪中,拖尾对成像的作用形式较为特殊,传统的扣除方案已不再适用,急需新的适合这一应用场合的方法。通过建立光谱仪中拖尾现象的数学模型,分析并仿真了常规拖尾扣除方法的失真问题,提出了适合推扫式光谱仪器的新方法。新方法使用已知图像信息估计景物的光谱分布,以此计算光学暗行通道所代表的拖尾信号在每个应用通道中的贡献,进而实现拖尾现象的扣除。遥感图像处理结果的较为理想的LSF表明,新方法效果良好,适用于推扫式成像光谱仪中CCD拖尾现象的扣除。     

小波域内的盲水印提取

在对水印嵌入位的假设检验基础上，提出了一种盲水印的逐位提取模型，并在小波域得到了实现.为平衡水印的不可见性和鲁棒性，根据图像小波域的特征，并利用上下文建模的方法为每个高频小波系数确定不同的水印幅度，从而使水印的嵌入强度依图像的特征而变化.由于水印提取的结果在很大程度上依赖于图像小波系数的统计分布模型，因此小波域中的系数建模采用了广义高斯分布，并使用极大似然法估计参量.实验表明，该方法具有良好的鲁棒性.     

基于伪魏格纳分布分解的合成孔径雷达图像目标与阴影分割

在二维魏格纳分布的框架内,针对魏格纳变换的交叉项问题和计算量大的问题,提出了合成孔径雷达图像局部伪魏格纳变换的目标和目标阴影的分割方法.首先,将合成孔径雷达图像进行二维伪魏格纳变换,得到各像素点的二维能量谱图|然后提取各像素点的二维能量谱图对应位置值形成多个不同频段的与原图像同大小的能量谱图|最后,对不同频段的能量谱图采用不同的处理方法后,将各能量谱图相加处理后形成区域标识图像,最终得到原图像的目标和目标阴影分割图像.本文利用该方法对MSTAR切片图像进行了分割试验,并对分割图像与频谱最大值距离或方位分割算法和基于双参量CFAR与隐马尔科夫联合分割算法进行了分割图像对比度对比.实验结果表明,采用本文算法的合成孔径雷达分割图像,对比度明显提高,且保留了目标图像细节.     

Symmetric semistable measures on Rn and symmetric semistable distribution operators in quantum mechanics

The paper contains two theorems. The first one deals with the classic probability, provides the form of the characteristic function of a symmetric, semistable probability distribution in Rⁿ. By a semistable probability distribution we mean a certain class of limit distributions of the sequence of random variables, containing the classic stable probability distribution and included in a set of infinitely divisible distributions.The result will be applied in the proof of the second theorem on non-commutative probability, in which we deal with the generalization of the central limit theorem in quantum mechanics.In the paper we define the symmetric semistable probability distributions for pairs of canonical observables in a state ?₀. In the second theorem we indicate that the symmetric semistable distribution operator is a ground state. By a ground state we mean a state for which Heisenberg's inequality becomes an equality.     

Optimal Randomness for Stochastic Configuration Network (SCN) with Heavy-Tailed Distributions

Stochastic Configuration Network (SCN) has a powerful capability for regression and classification analysis. Traditionally, it is quite challenging to correctly determine an appropriate architecture for a neural network so that the trained model can achieve excellent performance for both learning and generalization. Compared with the known randomized learning algorithms for single hidden layer feed-forward neural networks, such as Randomized Radial Basis Function (RBF) Networks and Random Vector Functional-link (RVFL), the SCN randomly assigns the input weights and biases of the hidden nodes in a supervisory mechanism. Since the parameters in the hidden layers are randomly generated in uniform distribution, hypothetically, there is optimal randomness. Heavy-tailed distribution has shown optimal randomness in an unknown environment for finding some targets. Therefore, in this research, the authors used heavy-tailed distributions to randomly initialize weights and biases to see if the new SCN models can achieve better performance than the original SCN. Heavy-tailed distributions, such as Lévy distribution, Cauchy distribution, and Weibull distribution, have been used. Since some mixed distributions show heavy-tailed properties, the mixed Gaussian and Laplace distributions were also studied in this research work. Experimental results showed improved performance for SCN with heavy-tailed distributions. For the regression model, SCN-Lévy, SCN-Mixture, SCN-Cauchy, and SCN-Weibull used less hidden nodes to achieve similar performance with SCN. For the classification model, SCN-Mixture, SCN-Lévy, and SCN-Cauchy have higher test accuracy of 91.5%, 91.7% and 92.4%, respectively. Both are higher than the test accuracy of the original SCN.     

A Noisy SAR Image Fusion Method Based on NLM and GAN

The unavoidable noise often present in synthetic aperture radar (SAR) images, such as speckle noise, negatively impacts the subsequent processing of SAR images. Further, it is not easy to find an appropriate application for SAR images, given that the human visual system is sensitive to color and SAR images are gray. As a result, a noisy SAR image fusion method based on nonlocal matching and generative adversarial networks is presented in this paper. A nonlocal matching method is applied to processing source images into similar block groups in the pre-processing step. Then, adversarial networks are employed to generate a final noise-free fused SAR image block, where the generator aims to generate a noise-free SAR image block with color information, and the discriminator tries to increase the spatial resolution of the generated image block. This step ensures that the fused image block contains high resolution and color information at the same time. Finally, a fused image can be obtained by aggregating all the image blocks. By extensive comparative experiments on the SEN1–2 datasets and source images, it can be found that the proposed method not only has better fusion results but is also robust to image noise, indicating the superiority of the proposed noisy SAR image fusion method over the state-of-the-art methods.     

Strength distributions and statistical spectroscopy. I. General theory

The strength distribution for an arbitrary excitation is given in terms of a double expansion, and its sum rules by single expansions, in polynomials defined by the initial and final energy spectra. In model spaces which are not too large, a rapid convergence, to within fluctuations, is assured by the action of a central limit theorem, as is shown in particular by considering the response of the system under infinitesimal deformations of the Hamiltonian. When larger spaces are decomposed into subspaces defined by a partitioning of the single-particle space a similar convergence results. At the same time, close contact is made with, and important corrections are found to, intuitive procedures which are often used for approximating strength distributions. The general features of the distribution are often easily understood in termsof a simple geometry made effective in the model space by the central limit theorem, and further features by exploiting the connection of this geometry to the unitary group of transformations in the single-particle space. Extensions are given for multipole strengths and sum rules, appropriate when the angular momenta (and isospins) are specified for the states involved in the transitions. Measures for the RMS fluctuations in the sum-rule quantities, and correlations between them, are given by combining the low-order-polynomial (statistically smoothed) strengths with an assumed Porter-Thomas distribution for the (high-order) strength fluctuations.     

Color fusion of SAR and FLIR images using a natural color transfer technique

Fusion of synthetic aperture radar (SAR) and forward looking infrared (FLIR) images is an important subject for aerospace and sensor surveillance. This paper presents a scheme to achieve a natural color image based on the contours feature of SAR and the target region feature of FLIR so that the overall scene recognition and situational awareness can be improved. The SAR and FLIR images are first decomposed into steerable pyramids, and the contour maps in the SAR image and the region maps in the FLIR image are calculated. The contour and region features are fused at each level of the steerable pyramids. A color image is then formed by transferring daytime color to the monochromic image by using the natural color transfer technique. Experimental results show that the proposed method is effective in providing a color fusion of SAR and FLIR images.     

Fluctuation Theory for a Three-Dimensional Model of Maxwellian Molecules

Under suitable assumptions, a functional central limit theorem is obtained for a three-dimensional model of Maxwellian molecules. This model is related to a nonlinear Boltzmann-type equation. It will be proved that the family of the distributions induced by fluctuation processes converges weakly.     

Probing quark-distribution amplitudes through generalized parton distributions at large momentum transfer

In the large momentum transfer limit, generalized parton distributions can be calculated through a QCD factorization theorem which involves perturbatively calculable hard kernels and light-cone parton distribution amplitudes of hadrons. We illustrate this through the H(q)(x,xi,t) distribution for the pion and the proton, presenting the hard kernels at leading order. As a result, experimental data on the generalized parton distributions in this regime can be used to determine the functional form of the parton distribution amplitudes which has thus far been quite challenging to obtain. Our result can also be used as a constraint in phenomenological generalized parton distribution parametrizations.     

Distribution of backscattered intensity in the distorted Born approximation: Application to C-band SAR images of woodland

Abstract

The intensity distribution in synthetic aperture radar (SAR) images of woodland is known to depend upon imaging conditions. Whilst phenomenological models can be used to match observed backscatter distributions, a physical model is needed to explain their origins. Images of woodland obtained during airborne SAR trials are analysed and shown to exhibit non-exponential intensity distributions. Expressions are derived for the moments of the intensity distribution using discrete scattering models based on the Born and distorted Born approximations. The predictions of the Born approximation are such that, at all but extremely high resolutions, the intensity statistics reflect only fluctuations in the number of discrete scatterers in resolution cells. In the distorted Born approximation it is revealed that, even at modest resolutions, fluctuations in both number and cross section of objects can influence intensity distributions. This is shown to be a direct consequence of the incorporation of attenuation effects in the distorted Born model. The theory is applied to scattering from a model woodland canopy and shown to yield intensity moments in close agreement with observations. The consequences of the model for other scattering situations are discussed.