基于背景图像差分的运动人体检测

在背景图像差分的基础上,提出一种改进的实时背景更新算法.通过背景学习过程,利用若干帧连续图像中每点像素灰度值样本,估算对应像素点的灰度值均方差.利用正态分布的"3σ原则"对图像背景进行实时更新,消除背景变化对人体分割的影响.提出基于图像平均灰度值的阈值分割算法,以提高算法对不同光照强度的适应能力.提出基于垂直投影图的阴影消除算法,分析表明,人体躯干和影子区的垂直投影特性不同,由此可以有效消除人体影子.实验结果表明,本算法在不同的光照情况下能够完整分割运动人体图像.     

一种采用背景统计技术的视频对象分割算法

利用背景统计技术从累积的帧差信息中构建出完整、可靠的背景区域,并将其与当前帧相比较,得到初始对象分割掩膜;再对之进行后处理,以消除噪声影响和平滑对象边界,从而获得较好的对象分割掩膜,并提取出视频对象。该算法不需要预知运动对象的形状、数目等,就能较好地从静止背景中分离出目标,实验证明,它具有一定的实用性和鲁棒性。     

基于运动一致性的视频对象分割方法研究

视频运动对象检测和分割是图像处理中最具挑战性的问题之一。针对目前大部分分割算法相当复杂而且计算量大的问题，提出了一种基于运动一致性的视频对象分割方法。该方法从MPEG压缩码流中提取运动矢量场来分割视频对象，首先对运动矢量场进行滤波和校正，然后进行全局运动补偿得到对象的绝对运动矢量场，最后采用K-means聚类算法对运动矢量场进行聚类分析从而分割出感兴趣的视频运动对象。MPEG标准测试序列的试验结果证明，该方法是有效的。     

基于高斯粒子滤波的当前统计模型跟踪算法

对于非线性系统估计问题,高斯粒子滤波器可以获得近似最优解,与粒子滤波器相比其优点是不需要重采样步骤和不存在粒子退化现象.采用高斯粒子滤波代替当前模型自适应跟踪算法中的卡尔曼滤波,将高斯粒子滤波与当前统计模型的优点相结合,提出了一种新的当前统计模型自适应跟踪算法,用于非线性非高斯系统的机动目标跟踪.MonteCarlo仿真表明,该算法跟踪精度优于标准的交互多模型算法和当前统计模型自适应跟踪算法,实时性好于交互多模型粒子滤波算法.     

基于改进C-V模型的超声图像分割方法

图像分割是超声医学图像学中的难题之一。改进的Chan-Vese（C-V）法加入了约束符号距离函数的能量项，避免了演化时候的重新初始化。在改进C-V模型的基础上，首先借用分水岭中的思想，找到分割目标的近似轮廓，并以此轮廓生成符号距离函数，然后采用改进的C-V法进行超声图像分割。实验表明，改进的方法有更高的精准度和对多目标分割的能力。     

分水岭变换和统计区域合并的图像分割算法研究

提出了一种基于分水岭变换和统计区域合并的图像分割方法.该方法综合利用高斯低通滤波、分水岭变换和统计区域合并,先对原始图像提取分割标记,然后利用Meyer分水岭变换对标记分水岭进行分割,最后利用概率统计的方法对过分割区域进行合并.该算法通过调节尺度参数可以实现由粗到细(coarse-to-fine)的分割.实验结果表明,这种简单可行的算法在分割噪声图像时依然有良好的效果,具有较强的鲁棒性.     

应用统计信号处理和模糊数学的图像融合算法

针对多传感器图像在像素级上的融合问题,将模糊数学理论引入到图像融合模型。该模型假定理想的融合后的图像包含场景所有的信息;将它乘上一个模糊因子,再加上随机噪声,可用来描述某一个成像传感器中获得的场景图像;不同的传感器对应不同的模糊因子和噪声。在此基础上,提出了建立在非多尺度分解框架下的图像融合算法。它以各传感器获取的图像作为输入条件,应用统计信号处理中的EM算法,求出针对不同传感器的噪声参数和模糊因子,通过迭代估计出融合的图像。实验结果显示,该算法获得的融合图像的互信息和联合熵分别达到3．5079和24．732,均优于加权平均融合法、小波融合算法和Laplacian融合算法的融合质量。     

基于差异积累的视频运动对象自动分割

针对视频运动对象的自动分割,本文给出了一种基于差异积累的自动分割算法。与传统的基于运动信息变化检测方法不同,该算法通过累积的帧差信息构建出可靠的背景,与当前帧比较进而提取出视频运动对象。本文提出了一种增强的基于Otsu法的自适应阈值化方法,能更准确地对背景差图像进行阈值化分割,克服了传统Otsu法阈值化容易失效的问题。改进的基于区域生长的定位方法更能避免传统方法的误定位及重定位的问题。实验结果表明,本文算法具有较好的实时性、自适应性以及鲁棒性,可以较为可靠地建立背景模型并进行实时更新,适用于刚体或非刚体存在平缓的光照变化以及摄像头微抖动的视频运动对象的自动分割。     

基于多阈值单水平集方法的医学图像分割

针对Chan-Vese的无边界主动轮廓模型(CV模型)只能区分前景与背景的缺点,提出了一种基于多阈值单水平集的医学图像分割方法,并将此方法应用于微创手术的预处理中.由于医学图像结构复杂,具有器官轮廓多连接等特点,因此使用常规的水平集方法进行分割往往不能取得理想的效果,而该方法采用修改目标泛函的方式引入多类分割,具有多区域分割的特点,只需经过一次单水平集的迭代循环,即可将图像根据灰度不同划分为多个区域,具有精确、快速等优点.对不同的合成图像和医学图像的实验结果表明,该方法实现了快速精确的多区域分割,能很好地提取到医学图像中的骨骼轮廓,分割效果达到了预期水平.     

基于神经网络的红外图像实时去斑方法

在分析红外图像斑纹产生机理基础上,提出利用人眼视网膜自适应调节原理的去斑方法。去斑时,根据距离不同作用不同原则,对中心及其周围像素分别予以不同权重;将斑纹信息加入求解过程,能有效快速去斑;去斑后,采用基于阈值的直方图规定化方法,以弥补对比度可能降低情况。研制出的DSP去斑系统实现了实时去斑功能。     

序列图像差分关联域的复杂全域运动估计

本文提出了一种新的基于差分关联域的复杂全域运动估计算法,该算法通过选取图像中梯度大的区域,避免了块匹配算法的误匹配问题,使算法的准确性得到保障.而且该算法结合运动估计技术中的参数估计与非参数估计技术,扩展了块匹配算法的适用范围,使得对复杂的全域运动进行有效估计成为可能.理论分析及大量实验证明,此算法能有效地处理包含缩放、旋转等分量的复杂运动.而且本算法将复杂的全域运动分解为各基本运动,使系统可以根据运算量的限制自适应地选择最优的运动估计算法.     

Natural disaster risk analysis for critical infrastructure systems: An approach based on statistical learning theory

Probabilistic risk analysis has historically been developed for situations in which measured data about the overall reliability of a system are limited and expert knowledge is the best source of information available. There continue to be a number of important problem areas characterized by a lack of hard data. However, in other important problem areas the emergence of information technology has transformed the situation from one characterized by little data to one characterized by data overabundance. Natural disaster risk assessments for events impacting large-scale, critical infrastructure systems such as electric power distribution systems, transportation systems, water supply systems, and natural gas supply systems are important examples of problems characterized by data overabundance. There are often substantial amounts of information collected and archived about the behavior of these systems over time. Yet it can be difficult to effectively utilize these large data sets for risk assessment. Using this information for estimating the probability or consequences of system failure requires a different approach and analysis paradigm than risk analysis for data-poor systems does. Statistical learning theory, a diverse set of methods designed to draw inferences from large, complex data sets, can provide a basis for risk analysis for data-rich systems. This paper provides an overview of statistical learning theory methods and discusses their potential for greater use in risk analysis.     

全变分自适应图像去噪模型

通过分析三种主要变分去噪模型(调和、全变分以及广义全变分模型)的优缺点,提出了一种基于全变分的自适应图像去噪模型。该模型根据噪声图像的信噪比,采用高斯滤波器对图像进行预处理,克服了全变分模型引入的阶梯效应;利用图像中每一像素点的梯度信息,自适应选取去噪模型中决定扩散强弱的参数p(x,y),使接近边缘处平滑较弱,远离边缘处平滑较强。数值实验表明,本方法在去除噪声的同时保留了图像的细节信息,取得了很好的降噪性能,其峰值信噪比(PSNR)在高噪声水平下,较其他变分方法至少提高1.0dB左右。     

基于时变Kanai-Tajimi模型的爆炸地震动拟合

借鉴完全非平稳地震动的研究方法,将VLACHOS C提出的完全非平稳时变Kanai-Tajimi模型用于爆炸地震动的拟合。结合实际爆炸地震动对这两种模型进行分析验证,通过比较拟合爆炸地震动与实际爆炸地震动的加速度时间历程曲线、反应谱、强度包线、归一化累积能量以及累积穿零次数,从时域和频域两方面来分析拟合效果,通过研究表明,单峰时变Kanai-Tajimi模型更适合用于爆炸地震动的拟合。     

基于时变Kanai-Tajimi模型的爆炸地震动拟合

借鉴完全非平稳地震动的研究方法,将VLACHOS C提出的完全非平稳时变Kanai-Tajimi模型用于爆炸地震动的拟合。结合实际爆炸地震动对这两种模型进行分析验证,通过比较拟合爆炸地震动与实际爆炸地震动的加速度时间历程曲线、反应谱、强度包线、归一化累积能量以及累积穿零次数,从时域和频域两方面来分析拟合效果,通过研究表明,单峰时变Kanai-Tajimi模型更适合用于爆炸地震动的拟合。     

Influence of statistical and constraint loss size effects on cleavage fracture toughness in the transition - A model based analysis

A database derived from tests on specimens with a large range of ligament (b) and thickness (B) dimensions was systematically analyzed to evaluate constraint loss and statistical size effects on cleavage fracture toughness. The objectives were to: (1) decouple size effects related to constraint loss, mediated by b and B, from those arising from statistical effects, primarily associated with B; and, (2) develop procedures to transfer toughness data to different conditions of constraint and B. The toughness database for a Shoreham pressure vessel steel plate, tested at a common set of conditions, was described in a companion paper. Quantification of constraint loss was based on an independently calibrated 3D finite-element critical stress-area, σ^∗-[K_Jm/K_Jc], model. The measured toughness data, K_Jm, were first adjusted using computed [K_Jm/K_Jc] constraint loss factors to the corresponding values for small scale yielding conditions, K_Jc=K_Jm/[K_Jm/K_Jc]. The K_Jc were then statistically adjusted to a K_Jr for a reference B_r = 25.4 mm. The B adjustment was based on a critically stressed volume criterion, modified to account for a minimum toughness, K_min, consistent with modest modifications of the ASTM E 1921 Standard procedure. The combined σ^∗-[K_Jm/K_Jc]-K_min adjustment procedure was applied to the Shoreham b − B database, producing a homogeneous population of K_Jr data, generally within the expected scatter. The analysis suggests that: (1) there may be a maximum B beyond which statistical size effects diminish, and (2) constraint loss in the three-point bend specimens begins at a relatively low deformation level. A corresponding analysis, based on a Weibull stress, σ_w-[K_Jm/K_Jc]-K_min, adjustment procedure, yielded similar, but somewhat less satisfactory, results. The optimized adjustment procedure was also applied to other K_Jm data for the Shoreham plate from this study, as well as a large database taken from the literature. The population of 489K_Jr data points, covering an enormous range of specimen sizes, geometries and test temperatures, was found to be consistent with the same master curve T₀ = −84 °C derived from the b − B database. Thus, calibrated micromechanical models can be used to treat size and geometry effects on K_Jm, facilitating using small specimens and data transfer to predict the fracture limits of structures.     

基于单样本学习的多特征人体姿态模型识别研究

随着人机交互、虚拟现实等相关领域的发展,人体姿态识别已经成为热门研究课题.由于人体属于非刚性模型,具有时变性的特点,导致识别的准确性和鲁棒性不理想.本文基于KinectV2体感摄像头采集的骨骼信息,结合人体角度和距离特征,提出了一种基于单样本学习的模型匹配方法.首先,通过对采集的骨骼信息进行特征提取,计算关节点向量夹角...     

A region-segmentation combining surrogate model based on L-indicator and N-fold cross-validation technique

For complex engineering problems, for which the mathematical models may be linear, low-order nonlinear or even high-order nonlinear, surrogate models which have high adaptability and accuracy are required. This article develops a method for constructing a region-segmentation combining surrogate model. It is based on the idea that in the entire experimental domain, different local regions may present different characteristics (linearity, low-order nonlinearity and high-order nonlinearity), and the entire domain should be divided into several subregions to be approximated by different surrogates so as to achieve high prediction accuracy in local regions. The preferred models in each subregion then constitute a weight-average combining surrogate model. The investigations reveal that the new model not only is more adaptive to analytically unknown functions, but also gives more accurate predictions. The method has been applied to three benchmark problems and a practical engineering problem, and the results maintain validity.     

On the prediction of the fatigue life of thin-walled structures based on a statistical model

We present experimental results of fatigue tests of AMg6 aluminum alloy model specimens-plates with a single notch, subjected to fully reversed and asymmetric cyclic axial loading. The experimentally determined lives of model specimens are used as the results of the base experiment for predicting the fatigue strength characteristics of large-scale cylindrical shells with circumferential ring-shaped stress raisers loaded by axial forces of a cyclic nature. A numerical study has been performed of the statistical model adopted for predicting the fatigue life of such shells. __________ Translated from Problemy Prochnosti, No. 3, pp. 112–122, May–June, 2006.