方差矩阵

Main contents of the thesis include point and edge extraction based on least-squares template matching; space circle template matching and reconstruction; outlier detection and elimination with fundamental matrix and trifocal tensor; improvement of Homography-based calibration technique, 2D-DLT and bundle adjustment based non-metric camera calibration technique with planar-scenes, point-line hybrid photogrammetry and visual inspection.

本文的主要研究内容包括影像格网点与直线段的自动最小二乘模板匹配；基于物方直接解的空间圆匹配与重建；利用基本矩阵和三视张量进行观测值粗差的剔除；同形矩阵摄像机标定算法的改进；基于二维 DLT 和光束法平差的非量测数字摄像机标定；点、线混合摄影测量模型与视觉检测精度分析等。

The linearly constrained minimum variance algorithm based on the power inversion array and its iterative formula are deduced in the paper, which avoids the complexities of matrix inversion.

详细推导了基于功率倒置阵列的线性约束最小方差算法及其递推公式，避免了矩阵求逆的复杂运算。

The algorithm uses posterior information to modify model's noise variance and markov transition matrix, so as to make IMM have adaptive ability.

该算法采用后验信息修正模型的噪声方差和马尔可夫转移矩阵，使IMM具有自适应能力。

This approach was based on the NonSubsampled Contourlet Transform. The means and variables of NSCT coefficients matrix in different subbands and various directions were extracted to form the feature vectors which were regarded as indexes of texture images in image database. Two similarity functions were used to compute the similarity between images. A texture retrieval system based on query image was developed.

该方法利用非下采样Contourlet变换对图像进行分解，提取不同子带和不同方向变换系数矩阵的均值和方差为特征向量，作为数据库中纹理图像的索引，并利用两种不同的相似度函数计算图像之间的相似度，建立了一套基于示例查询图像的纹理图像检索系统。

In Chapter 1, we introduce some theories and methods about the estimation of regression coefficient and do some preparations including some theories in the mean vector, the covariance matrix, the quadratic form of random vector, matrix derivative.

第1章概括介绍线性回归模型回归系数估计的一些理论和方法，并给出均值向量与协方差阵、随机向量的二次型、矩阵微商等预备知识。

Methods Measure the T scores of pilots.Extract six common factors of P,N,I,M,A,F with major com ponent analysis and get the loading matrix and factor construction by varimax or thogonal rotation.

测查MMPI的飞行员T分值，采用主成分分析法提取P、N、I、M、A、F6个公因子进行方差极大旋转，得出因子负荷矩阵和因子结构状况。

Based on the concept of complete consistency for triangular fuzzy number complementary judgment matrix, a non-linear programming model is set up based on least variance method.

根据三角模糊数互补判断矩阵完全一致性的概念，建立了一个基于最小方差的非线性规划模型。

Coherent interference can be suppressed by the EMCMVB efficiently. The EMCMVB first gets the modified signal subspace by using the estimated arrival angles of coherent interferers and the eigenstructure of the correlation matrix, and then projects the MCMVB weight vector onto the modified signal subspace.

EMCMVB是利用估计得到的相干干扰方向和阵列信号相关矩阵特征分解来得到修改的信号子空间，再把多约束最小方差波束形成器的权矢量向修改的信号子空间投影来得出新的权矢量。

For the multisensor systems with uncorrelcted noises,and with unkonw noise varaies.substituting the information fusion noise variance setimators based on the correlct method into the centralized fusion optimal information filter,a self-turning centralized fusion information filter is presenced compared with the self-turning centralized fusion Kalman filter based on the Riccati equation,it can be avoid to compute zhe inverse matrix with high dimension,so that the calculation hurden can be reduced significantly.

对于带未知噪声系统和不相关噪声的多传感器随机系统，将基于相关方法得到的噪声方差带入到集中式融合最有信息滤波器，提出自校正集中式融合信息滤波器。同基于Riccatia方程的集中式融合Kalman滤波器相比它避免了计算高维矩阵的逆，从而减少了计算负担。

The nonlinear transformations called the unscented transformation is applied to propagate the means and covariances of states, and the errors caused by linearization of extended Kalman filter is eliminated.

UKF采用确定性采样策略，通过UT变换实现状态均值和方差的非线性传播，避免了扩展卡尔曼滤波产生的线性化误差，并且无需计算雅可比矩阵。

In the United States, chronic alcoholism and hepatitis C are the most common ones.

在美国，慢性酒精中毒，肝炎是最常见的。

If you have any questions, you can contact me anytime.

如果有任何问题，你可以随时联系我。