The ordering process , the integer ldlt factorization and the ambiguity screen are three parts of the osd . the osd is based on diminishing the condition number of covariance matrix Osd处理过程包括排序、整数ldl ~ t分解和模糊度筛选，该算法的去相关处理过程，是以降低协方差矩阵的条件数为准则。
The experimental results on orl face database show that by the proposed improved method , eigenvalues of the optimized covariance matrices become larger and the matrices inverses become more stable 在orl人脸库上的实验表明，优化后协方差矩阵的特征值均显著变大，使该矩阵的逆阵稳定性得到了提高。
At the same time , the computational complexity is decreased . improved music algorithm decomposes the signal covariance matrix and then make the signal subspace be orthogonal to the noise subspace by 改进music算法通过对接收信号协方差矩阵作预处理,使信号协方差矩阵分解后得到的信号子空间与噪声子空间能够正交。
Ensuring the short - range , medium - range and long - range errors , using the covariance evaluation and processing program for experimental data , the covariance matrices were constructed for each set of experimental data 在确定各家数据的短程、中程和长程误差的大小之后，用实验数据协方差评价处理程序，构造出每一家数据的协方差矩阵。
This method constructs covariance matrix by utilizing data vectors in different range lines and projects phase error vector into noise sub - space which is formed by eigendecomposing the covariance matrix 该方法利用不同距离单元的观测矢量构造协方差矩阵，然后通过对协方差矩阵特征分解得到噪声子空间，最后将相位误差矢量向噪声子空间投影来估计多普勒调频率。
Based on the multi - scale representation theory , we present a reduced order model for the solving of the inverse problem . also the relative error covariance matrix is used to analyze the performance of models with different orders ; 4 基于小波多尺度表示理论给出逆问题求解的多尺度降阶模型，同时用相对误差协方差矩阵对阶数不同的降阶模型的估计精度进行分析； 4
The main work is : 1 ) this text gives a study of the signal models of two kinds of distributed sources : coherent distributed sources and incoherent distributed sources , analyzes eigenvalue distribution of correlation matrix 主要工作如下： 1 ）研究了两种分布式目标信号模型：相干分布式目标模型和非相干分布式目标模型。并分析了分布式目标协方差矩阵的特征值分布。
They are detailed as follows : 1 . to construct covariance matrix with time correlation and distance correlation . both observational data and theoretical model results has its respective errors , upon which both covariance matrix being produced 具体内容为如下几个方面： 1建立时间相关的协方差矩阵处理方法在估值类的同化方法中，观测数据与理论模型结合的桥梁是各自的协方差矩阵。
Main contents include man - made targets extraction based on scattering matrix , man - made targets extraction based on covariance matrix [ c ] and the extraction based on geometric feature . first , this dissertation provides an overview of the polsar theory 主要内容包括:基于散射矩阵的polsar图像目标提取算法,基于协方差矩阵polsar的目标提取算法和基于几何特征的人造目标提取算法等方面。
One - dimension noise subspace - based method always assumes that only the minimum eigenvalue of the signal covariance matrix is the noise eigenvalue , and the corresponding eigenvector is the true noise vector and constructs the one - dimension noise subspace 基于一维噪声子空间方法始终认定只有信号协方差矩阵的最小特征值才是噪声特征值,其对应的特征向量才是真正的噪声向量,并构成一维噪声子空间。
