一种多视极化SAR广义K-分布的研究

本文提出一种极化合成孔径雷达(SAR)广义K-分布模型用于描述经多视处理的、高分辨率的全极化SAR成像数据的统计特性,其特点是各极化通道的纹理为部分相关的gamma分布.利用纹理与相干斑乘性模型,首先导出了部分相关纹理的联合概率密度函数和多视广义极化SARK-分布概率密度函数;然后给出了基于联合矩的分布参数估计方法;最后利用真实SAR数据进行了模型的试验.     

Beta分布下基于白化滤波的极化SAR图像海面舰船目标CFAR检测方法

在杂波纹理服从Beta分布的极化合成孔径雷达（Polarimetric Synthetic Aperture Radar,POLSAR）图像目标检测中,提出了一种基于多视极化白化滤波（Multilook Polarimetric Whitening Filter,MPWF）的恒虚警（Constant False Alarm Rate,CFAR）检测解析新方法.首先,假设乘积模型中纹理变量服从Beta分布,推导得到MPWF检测量的概率密度函数（Probability Density Function,PDF）.然后,对概率密度函数积分得到虚警概率关于检测门限的解析式,并设计相应的CFAR检测流程.最后,提出了基于MPWF的对数累积量估计方法,对Beta分布纹理变量参数u和v进行估计.通过实测数据验证了新方法的有效性.实验结果表明Beta分布对某些区域的极化SAR数据有更好的拟合效果,同时新方法与已有方法相比具有更好的CFAR保持能力.     

Wishart分布矩阵行列式值的统计特性及其在参数估计中的应用

 极化协方差矩阵行列式值表征了Pol-SAR图像目标散射极化散布程度,它在参数估计中表现出突出的优势,研究行列式值的物理内涵及其统计特性对揭示行列式值的优势的本质有着重要的理论价值.在此背景下,将表征目标散射极化散布程度的标准行列式值与表征目标散射随机性的极化熵、极化散度进行了比较,给出了它们与平面极化度和方向极化度的数学关系,指出他们在描述目标散射随机特性方面具有部分一致性,存在差异的本质在于三个随机性描述参数对极化方向随机度和平面极化度的加权方法的不同.然后研究了极化协方差矩阵行列式值的统计分布特性,给出了极化协方差矩阵行列式值的概率密度函数,指出在视图数和样本数足够的情况下,其行列式值的测量均值是渐进无偏估计.针对样本数较少时等效视图数参数估计误差较大的问题提出了最大似然修正算法,提高了小样本数目时的等效视图数参数估计精度.最后用仿真数据对理论推导进行了验证,结果表明了理论结果的正确性.极化协方差矩阵行列式值的统计特性和目标散射随机性参数的研究为目标极化散射随机特性、参数估计和目标检测等的研究提供了新思路.     

Fisher分布下基于白化滤波的极化SAR图像CFAR检测方法

在已有的极化合成孔径雷达（SAR）图像恒虚警（CFAR）检测方法中,存在着高分辨下杂波模型适用性差的难题.为此提出了一种Fisher分布下具有虚警概率解析表达形式的CFAR检测方法.首先,在乘积模型框架下,引入Fisher纹理变量,推导出了极化白化滤波（PWF）检测量的概率密度函数（PDF）.然后,对PDF积分得到了虚警概率关于检测门限的解析表达形式,并设计了相应的CFAR检测算法流程.最后,通过机载合成孔径雷达（AIRSAR）实测数据比较了新方法和双参数恒虚警（2P-CFAR）算法及已有的基于K分布、G₀分布、Wishart分布的CFAR检测方法的检测性能.结果表明新方法能有效检测出目标,且鲁棒性较强,相比于其他检测方法,品质因数平均高出32.66%.     

基于矩方法的多视广义极化K-分布参数的估计

本文提出通过归一化极化联合矩对SAR多视广义极化K-分布参数进行估计的方法。本文深入阐述了SAR各个极化通道之间的纹理相关性;在此基础上给出了多视广义极化K-分布波形因子,纹理相关系数和相干斑相关系数的偏差性能分析;本文最后进行了蒙特卡洛实验论证,进一步证明了该方法的良好估计性能。     

Fisher纹理分布下基于匹配滤波的极化SAR图像CFAR检测方法

在已有的极化合成孔径雷达（PolSAR）图像恒虚警（CFAR）检测方法中,存在着高分辨下杂波模型适用性差的难题。为此提出了一种Fisher分布下的CFAR检测方法,并定义虚警损失率（CFAR Loss,C_L）以量化评估算法的恒虚警保持性能.首先,在乘积模型框架下引入Fisher纹理变量,推导出了多视极化匹配滤波（Multi-look Polarization Matched Filter,MPMF）检测量的概率密度函数（PDF）.然后,对PDF积分得到了虚警概率的闭合解析式,并设计了CFAR检测流程.仿真数据和机载合成孔径雷达（Airborne SAR,AIRSAR）数据实验结果表明,与基于K分布、G₀分布、Wishart分布的CFAR检测算法以及双参数恒虚警（two-Parameter CFAR,2P-CFAR）算法相比,新方法具有良好的恒虚警保持性能和检测性能,具有较强的鲁棒性,且运算时间未明显增加,相比于其他检测方法,品质因数（Figure of Merit,FoM）平均高出12.80%.     

基于小波变换和马尔可夫随机场的极化SAR图像自动分类

本文提出一种对极化合成孔径雷达(SAR)图像进行自动多分辨率分类的方法。首先利用多视极化白化滤波(MPWF)抑制极化SAR图像的相干斑,得到反映地物辐射特征的纹理SAR图像,然后利用小波变换(WT)提取不同分辨率的纹理信息,在最低分辨率级利用Akaik信息准则(AIC)自动估计图像中的纹理类数,进而在各个分辨率级利用马尔可夫随机场(MRF)模型表征各像素间的空间关联信息,并分别利用最大似然(ML)方法和循环条件模式(ICM)进行自动的模型参数估计和最大后验概率(MAP)分类,最后应用NASA/JPL机载L波段极化SAR数据验证了本文所提分类方法的有效性和优越性。     

多视SAR图像K分布特征函数与统计量研究

K分布是一种广泛应用于描述SAR图像杂波统计特性的概率模型,然而对于其特征函数及各类统计量的研究却少见于报道。本文基于Fourier变换和Mellin变换,详细推导了多视K分布第一类和第二类第一、第二特征函数的解析表达式,获得了原点矩、累积量、对数矩和对数累计量的计算公式,在此基础上给出了基于原点矩的参数估计器和基于对数累积量的参数估计器。最后利用Kendall近似展开法,理论分析了两种估计器的正规化偏差与方差。结果表明,对于小样本不均匀SAR图像场景,对数累积量估计器的性能明显优于矩估计器。      

基于混合矩的极化SAR图像K分布模型参数估计新方法

K分布模型在极化合成孔径雷达(PolSAR)图像建模领域中获得广泛应用。其模型参数估计的精度将直接影响到模型拟合的准确性。目前普遍采用的K分布参数估计方法是基于协方差矩阵Mellin变换的对数累积量的估计方法。但是该方法没有解析的表达式,数值计算运算时间较长,另外在形状参数1时估计偏差较大。为此该文提出一种基于|z|rlg|z|混合矩的参数估计新方法,该方法对不同形状参数值下的参数估计具有较好的适应性,并且在值较小时估计性能优于对数累积量方法。同时在r=1/d时该方法有解析的表达式,其运算时间优于对数累积量方法。最后用仿真数据和实测数据对新方法和已有参数估计方法的结果进行了比较,验证了基于混合矩估计方法的准确性与有效性。该方法为PolSAR图像统计模型参数的快速有效估计提供了新手段。     

一种基于非参数密度估计和马尔可夫上下文的SAR图像分割算法

在研究传统的基于参数的合成孔径雷达(SAR)图像统计模型基础上,为了精确估计高分辨率SAR图像的统计分布,该文提出了一种结合基于核函数的非参数估计和马尔可夫上下文的SAR图像分割算法。该算法首先采用基于核函数的非参数方法估计SAR图像的统计分布,然后将此统计量作为图像分割的似然函数,利用马尔可夫上下文约束进行SAR图像分割。该文通过软件仿真对新算法和基于参数的统计模型的算法的效果进行了比较。研究发现,基于核函数的非参数估计方法仅仅依赖实际数据,在无法准确获取分布函数解析式的情况下往往具有更好的效果。实验证明,基于核函数的非参数估计方法对高分辨率SAR图像中较为复杂的场景如城区的提取取得了更为满意的结果。     

A Birnbaum-Saunders accelerated life model

The 2-parameter family of probability distributions introduced by Birnbaum and Saunders characterizes the fatigue failure of materials subjected to cyclic stresses and strains. It is shown that the methods of accelerated life testing are applicable to the Birnbaum-Saunders distribution for analyzing accelerated lifetime data, and the (inverse) power law model is used due to its justification for describing accelerated fatigue failure in metals. This paper develops the (inverse) power law accelerated form of the Birnbaum-Saunders distribution, and explores the corresponding inference procedures-including parameter estimation techniques and the derivation of the s-expected Fisher information matrix. The model approach in this paper is different from an earlier work, which considered a log-linear form of a model with applications to accelerated life testing. Here, using an example data set, the fitted model is effectively used to estimate lower distribution percentiles and mean failure times for particular values of the acceleration variable. The benefits of having an operable closed form of the Fisher information matrix, which is unique to this article for this model, include interval estimation of model parameters and LCB on percentiles using relatively simple computational procedures     

The bivariate generalized-Κ (ΚG) distribution and its application to diversity receivers

The correlated bivariate generalized-K (K_G) distribution, with not necessarily identical shaping and scaling parameters, is introduced and studied. This composite distribution is convenient for modeling multipath/shadowing correlated fading environments when the correlations between the signal envelopes and their powers are different. Generic infinite series expressions are derived for the probability density function (PDF), the cumulative distribution function (CDF) and the joint moments. Assuming identical shaping parameters, simpler expressions for the PDF, CDF and the characteristic function (CF) are provided, while the joint moments are derived in closed form. Furthermore, the PDFs of the product and ratio of two correlated K_G random variables are obtained. Capitalizing on these theoretical expressions for the statistical characteristics of the correlated K_G distribution, the performance analysis of various diversity reception techniques, such as maximal ratio combining (MRC), equal gain combining (EGC) and selection diversity (SD), over bivariate K_G fading channels is presented. For the SD, the outage probability is studied, while for the MRC and EGC the average bit error probability is obtained. The proposed analysis is accompanied by numerical results, clearly demonstrating the usefulness of the theoretical approach as well as the appropriateness of the K_G distribution to model multipath/shadowing fading channels.     

On estimating parameters in a discrete Weibull distribution

Two discrete Weibull distributions are discussed, and a simple method is presented to estimate the parameters for one of them. Simulation results are given to compare this method with the method of moments. The estimates obtained by the two methods appear to have almost similar properties. The discrete Weibull data arise in reliability problems when the observed variable is discrete. The modeling of such a random phenomenon has already been accomplished. Estimation of parameters in these models is considered. Since the usual methods of estimation are not easy to apply, a simple method is suggested to estimate the unknown parameters. The estimates obtained by this method are comparable to those obtained by the method of moments. The method can be applied in most inferential problems. Though the authors have restricted themselves to type I distribution, their method of proportions for the estimation of parameters can be easily applied to the type II distribution as well     

Distribution of Path Loss for Wireless Personal Networks Operating in a Square Region

Path loss plays fundamental roles in system design, spectrum management, and performance evaluation. The traditional path loss model has a slight inconvenience; it depends on the unknown distance. In this letter, we explore the probability distribution function (PDF) of path loss in an indoor office environment by randomizing out the distance variable. It is shown that the resulting PDF is not Gaussian‐like but is skewed to the right, and both the PDF and the moments are related to the size of the office instead of the unknown distance. To be specific, we incorporate the IEEE 802.15.4a channel parameters into our model and tabulate the cumulative distribution function with respect to different room sizes. Through a simple example, we show how our model helps a cognitive spectrum user to infer path loss information of primary users without necessarily knowing their transmitter‐receiver distance.     

Reconciling statistical models with practical experience of reverberation chambers

Presented are novel theoretical probability density functions (PDF) for the magnitude and phase of electromagnetic fields inside reverberation chambers (RC) operating in a `good-but-imperfect regime?. The derivation is based on considering two Gaussian random variables with mean values, variances and correlation between them that depart from the ideal assumptions. A multivariate approach using a complete joint Gaussian distribution of these variables is defined. Marginal distributions obtained by integration of this twodimensional joint PDF are compared with theoretical PDFs for ideal situations, and significantly lower rejection rates are experienced for field data measured in real RCs. Additionally, these novel marginal PDFs are highly general since they are able to describe both ideal and non-ideal stirred fields.     

Hybrid nonlinear moments in array processing and spectrum analysis

The aim of this paper is to provide a theoretical basis for the use of hybrid nonlinear (HNL) moments in array processing and spectrum analysis. These moments are defined as the expected value of the product of one random variable times a nonlinear function of another random variable. They generalize a class of twofold higher order moments, and their additional flexibility can be exploited for optimization purposes or for computational convenience. A number of properties beyond the classical Bussgang's (1952) and Price's (1958) theorems are found for HNL moments and matrices, making these statistics suitable for harmonic analysis and bearing estimation. Covariance based and higher order moments based methods are extended to the HNL moments domain, and a new class of Gaussian noise rejecting statistics is added to cumulants. The properties of some classes of matrices of HNL moments of practical interest are analyzed in detail     

基于海杂波分形特性的目标检测新方法

为克服海杂波分形维数值估计需要较长的数据长度的缺陷,提出了基于扫描间积累的分形维数值估计方法.首先研究了几种常用的分维值估计方法,从分维值的定义出发证明了分段分维估计方法与连续数据估计方法的等价性,然后通过迭代算法求得分形维数估计值.仿真结果表明提出的方法降低了对雷达驻留时间的要求,并且具有较小的运算量,提高了微弱目标的检测性能.该方法可以用于岸基对海雷达检测海表面微弱目标.