基于能量信息耦合梯度调节机制的图像修复算法设计与应用

为了克服当前图像修复算法主要依靠图像的置信度信息来获取优先修复块,忽略了图像的能量信息,导致修复结果中存在不连续及伪吉布斯现象等缺陷,设计了基于能量信息与梯度调节机制的图像修复算法;首先,通过区域能量函数来求取图像的能量信息,以计算待修复块的优先权信息,得到优先修复块;然后,基于图像梯度模值,建立梯度调节机制,以调节样本块的大小,获取与图像纹理相适应的样本块尺寸;引入平方差求和(SSD)函数,以确定最优匹配块;最后,通过像素点间的差异性,构造相似惩罚因子,以更新置信度项,完成图像的修复;实验结果显示,在所提算法中将变化阈值取值为0.5时,较当前图像修复方案而言,所提算法具备更好的修复性能,所得到的修复图像拥有更好的纹理连贯性与更高的结构相似(SSIM)值;当对像素丢失比例达到50%的图像进行修复时,较对照组算法而言,所提算法修复图像的SSIM值分别提高了6.5%和17.5%。

一种改进的全变分自适应图像去噪模型

目的比较几种不同的变分图像去噪模型的优缺点,提出一种新的根据局部梯度信息自适应调整光滑性测度的模型。方法新模型增加了梯度模值与两个门限的比较,小于门限1采用各向同性扩散去噪模型,大于门限2采用TV去噪模型,梯度模值在两个门限之间的采用自适应去噪模型。结果新模型在强噪声水平下,改善了人眼主观视觉感受,均方误差(MSE)降低了约35,峰值信噪比(PSNR)提高了约2.25dB。结论新模型改进了原有模型,进一步减少了"阶梯"效应,得到了更好的去噪效果。

结合颜色和梯度信息的稀疏图像修复算法

现有基于稀疏性的图像修复算法仅利用颜色信息衡量样本块的相似度,易降低修复区域内结构部分的连通性及与邻域信息的连续一致性,同时在全局范围内搜索匹配块也增加了算法的运行时间.为解决上述问题,利用颜色与梯度模值信息度量样本块之间的距离,构造新的相似度以确定块结构稀疏度函数,利用块结构稀疏度确定填充顺序,同时构造新的匹配准则函数寻找匹配块;并利用块结构稀疏度值能够较好地反映样本块所处区域特征的特性,根据块结构稀疏度值自适应确定局部搜索区域大小.并通过实验验证在不同图像中颜色信息与梯度信息所占比例不同.实验结果表明,该算法较对比算法能够更好地保持结构部分的连贯性及与邻域信息的连续一致性,在峰值信噪比上至少提高1dB,并且算法速度提高4~11倍.

基于GAC模型实现交互式图像分割的改进算法

提出了一种改进的交互式图像分割算法。采用全变分去噪模型对图像进行预处理,在去除噪声的同时更好地保护了边缘;提出了一种对梯度模值进行曲率加权的边缘检测方法,采用该方法获得图像的边缘点集;将边缘点集中曲率较大的边缘点作为候选边界点推荐给用户;用户通过主观判断,在候选边界点中选择合适的"初始边界点",算法便可采用GAC模型完成对目标的分割。实验结果表明,改进算法提高了交互式图像分割的自动化程度,有效地减少了交互过程中的人工参与量。

基于全向小波的图像边缘检测算法

针对现有边缘检测算法难以提取图像任意方向的边缘特征,提出基于全向小波的图像边缘检测算法.首先,定义了全向小波的概念、构造其模型并推导了全向小波的最大值与梯度模值相等的关系.理论分析表明本文算法始终沿小波变换值的最大值方向提取边缘.然后,选择二维高斯函数实例化模型,以8方向和3×3变换窗为例进行算法设计.标准图像对比试验表明本文算法能提取更多方向的边缘特征、边界清晰度也比SADD算法、Canny算子分别高出约2.17%、8.66%.

改进的Canny算法及其在图像编码中的应用

分析仅以梯度模值作为依据的Canny边缘检测算法的不足,提出一种以曲率加权梯度模值为判据的改进算法以满足基于边缘信息的图像编码压缩对边缘提取的严格要求。实验结果表明,改进的Canny算法能更有效地抑制对于表征图像对象无重要意义的杂乱边缘,更好地反映图像的结构特征,性能优于现行的Canny算法。

基于方差调节策略耦合结构特征的图像修复算法

针对当前图像修复算法主要采用图像的R、G、B信息来获取最优匹配块,忽略了图像的结构特征,导致修复图像中含有纹理不连续及块现象等问题,设计了方差调节策略耦合结构特征的图像修复算法。首先,借助图像的梯度模值来构造结构度量因子,以度量图像的结构特征。联合置信度项、结构度量因子和数据项,构造优先权函数,以寻求优先修复块。然后,采用图像的方差特征,建立方差调节策略,考虑图像纹理的动态变化,寻求与当前纹理情况最吻合的样本块尺寸。最后,将结构度量因子引入到最优匹配块的搜索过程中,以弥补通过R、G、B信息搜索最优匹配块时所忽略图像结构特征的不足,得出最优匹配块修复图像,实现损坏区域的复原。通过实验结果发现,较当前修复方案而言,所提算法的修复图像具备更优的纹理连续性。

基于偏微分方程的误差分散逆半调改进算法

结合误差分散半调噪声特征,提出一种基于偏微分方程的逆半调改进算法。通过研究偏微分方程的去噪原理,以3次B样条函数作为扩散函数,采用迭代求解偏微分方程的方法估计逆半调图像,计算每次迭代前后图像梯度模值的增量以确定平滑度的调节参数,解决偏微分方程在应用中的参数选择自适应问题。实验结果表明,该算法在图像整体平滑度和细节保持能力上都具有较好的效果。

基于MIL的图像分割算法研究

传统分水岭图像分割算法主要基于梯度模值图像,寻找集水盆地,易产生过分割现象。本文提出了基于MIL改进的分水岭图像分割算法,借助MIL软件开发性强、函数库功能强大等特点,对粘连目标图像进行了分割。实验证明,该方法具有较强的有效性和实用性。

基于多维传输函数的三维医学图像系统研究

传输函数是直接体绘制的基础,传输函数的设计是体视化结果的关键。本文在传输函数的定义上,讨论了将体数据标量值和梯度模值作为定义域的多维传输函数的原理及实现方法。针对传输函数与位置无关,难以提取感兴趣的区域(Region Of Interest简称ROI),提出了基于三切面交互式体数据的切割方法,从而有效的提取感兴趣区域。最后,设计和实现了一个基于多维传输函数的三维医学图像系统,获得了很好的体绘制效果。

Numerical simulation on size effect of copper with nano-scale twins

To understand the tensile deformation of electro-deposited Cu with nano-scale twins, a numerical study was carried out based on a conventional theory of mechanism-based strain gradient plasticity （CMSG）. The concept of twin lamella strengthening zone was used in terms of the cohesive interface model to simulate grain-boundary sliding and separation. The model included a number of material parameters, such as grain size, elastic modulus, plastic strain hardening exponent, initial yield stress, as well as twin lamellar distribution, which may contribute to size effects of twin layers in Cu polycrystalline. The results provide information to understand the mechanical behaviors of Cu with nano-scale growth twins.

Investigation of Barree-Conway non-Darcy flow effects on coalbed methane production

Coalbed gas non-Darcy flow has been observed in high permeable fracture systems,and some mathematical and numerical models have been proposed to study the effects of non-Darcy flow using Forchheimer non-Darcy model.However,experimental results show that the assumption of a constant Forchheimer factor may cause some limitations in using Forchheimer model to describe non-Darcy flow in porous media.In order to investigate the effects of non-Darcy flow on coalbed methane production,this work presents a more general coalbed gas non-Darcy flow model according to Barree-Conway equation,which could describe the entire range of relationships between flow velocity and pressure gradient from low to high flow velocity.An expanded mixed finite element method is introduced to solve the coalbed gas non-Darcy flow model,in which the gas pressure and velocity can be approximated simultaneously.Error estimate results indicate that pressure and velocity could achieve first-order convergence rate.Non-Darcy simulation results indicate that the non-Darcy effect is significant in the zone near the wellbore,and with the distance from the wellbore increasing,the non-Darcy effect becomes weak gradually.From simulation results,we have also found that the non-Darcy effect is more significant at a lower bottom-hole pressure,and the gas production from non-Darcy flow is lower than the production from Darcy flow under the same permeable condition.

Numerical analysis and experimental study of hydrogen production from dimethyl ether steam reforming

An experimental and theoretical study of steam reforming of dimethyl ether was carried out in a processor for fuel cell vehicles to explore the effect of temperature gradient and hydrogen content of the processor.A steady-state,laminar,two-dimensional axi-symmetric model was proposed to investigate the fluid flow,heat transfer and chemical reactions in the dimethyl ether steam reforming processor using porous medium approach.The numerical model was established with Star-CD program using SIMPLE algorithm and finite volume method.Experimental verification of the two-dimensional mathematical model was conducted.The numerical results coincided well with the experimental data.The effects of the parameters on the temperature gradient and hydrogen content of the processor were studied using the numerical model.

Flow Instability of a Centrifugal Pump Determined Using the Energy Gradient Method

The stability of the centrifugal pump has not been well revealed because of the complexity of internal flow. To analyze the flow characteristics of a centrifugal pump operating at low capacity, methods of numerical simulation and experimental research were adopted in this paper. Characteristics of the inner flow were obtained. Standard k-s turbulence models were used to calculate the inner flow of the pump under off-design conditions. The distri- bution of the energy gradient function K was obtained by three-dimensional numerical simulation at different flow rates. The relative velocity component was acquired from the absolute velocity obtained in particle image velocimetry. By comparing with experimental results, it was found that flow instability occurs at the position of maximum K. The flow stability reduces with an increasing flow rate. The research results provide a theoretical basis for the optimization design of a centrifugal pump.

Study on the spectral response of fiber Bragg grating sensor under non-uniform strain distribution in structural health monitoring

Many theoretical studies have been developed to study the spectral response of a fiber Bragg grating (FBG) under non-uniform strain distribution along the length of FBG in recent years. However, almost no experiments were designed to obtain the evolution of the spectrum when a FBG is subjected to non-uniform strain. In this paper, the spectral responses of a FBG under non-uniform strain distributions are given and a numerical simulation based on the Runge-Kutta method is introduced to investigate the responses of the FBG under some typical non-uniform transverse strain fields, including both linear strain gradient and quadratic strain field. Experiment is carried out by using loads applied at different locations near the FBG. Good agreements between experimental results and numerical simulations are obtained.

Sub-ensemble study of pressure gradient effects on the variation of Karman constant

Using data from direct numerical simulation （DNS） of incompressible and compressible channel flow, we develop a method of sub-ensemble decomposition to investigate the pressure gradient effect on the Karman constant and the additive constant B characterizing the mean velocity profile （MVP）. The sub-ensemble decomposition is defined according to the magnitude of vertical fluctuation velocity, which mimics coherent motions like ejection and sweep. DNS data analysis shows that each sub-ensemble displays a distinct Karman constant, with a variation which mimics effects of pressure gradient. The latter is demonstrated by a relation between sub-ensembles＇ km and Bm similar to empirical data under various pressure gradients. A set of global parameters, k0-pg=0.39 ＆ B0-pg=5.5, are then derived for interpreting two constants observed by Nagib et al.