基于最稳定极值区域方法的目标检测与跟踪

针对非刚体目标的精确实时跟踪问题,提出了一种融合先验形状信患的基于最稳定极值区域(MSER)检测器的跟踪算法.首先,利用训练样本建立目标颜色特征的混合模型,生成目标统计颜色概率图,为最大稳定区域方法提供概率统计依据.其次,利用基于最稳定极值区域方法给出最稳定的分割结果.最后,利用训练样本得到目标的先验动态形状模型,并且融合目标形状信息与通过MSER算法生成的稳定区域信息,去除虚假分割结果,提高目标检测精度与跟踪性能.实验结果证明,该算法能在视频序列图像中有效检测并跟踪目标.     

基于先验形状的CV模型肝脏CT图像分割

针对目标被部分遮挡或部分信息丢失情况下CV模型不能正确识别的问题,提出一种新的分割算法。首先,利用数学形态学对原肝脏图像进行滤波,并结合其他算法建立肝脏先验形状;然后,采用边缘查找和区域标定等算法,对肝脏先验性状的边缘以及边缘内外区域进行赋值,构建执行效率高的符号函数距离函数,将其通过形状比较函数嵌入到CV模型的能量泛函中,形成新的基于先验形状的CV模型,并将此模型用于分割存在干扰或者被部分遮挡的肝脏CT图像。与CV模型分割结果相比,本文算法能在目标周围存在干扰信息或者被部分遮挡的情形下,成功地正确识别出目标区域。     

结合形状先验的图割目标分割方法

为了减少图像目标分割过程中噪声、阴影、背景复杂等因素的影响,将形状先验引入图割图像分割框架中,提出一种结合形状先验的图割目标分割方法。该方法给出了形状先验的定义,并将其转化为势函数的形式加入到能量函数中,通过能量最小化过程得到最终分割结果。同时,图像对齐过程能够适应形状模版与待分割目标之间的仿射不同。形状先验能够较好地约束目标边界,相比与传统图割算法,分割结果有了明显改善。实验结果表明,该算法具有有效性。     

浅析GVF模型在图像分割中的改进应用

本文在分析GVF模型的变形过程中,基于其存在的问题,提出了将傅里叶算子与GVF模型相结合的改进算法.这种算法将形状先验知识引入图像的分割过程中,利用它进行形状相似性的测量.该算法可以根据待提取目标的形状特点,不断自动调整蛇模型的参数,直至提取出合适的目标边缘,在一定程度上解决了图像中目标边缘难以提取的问题.     

Faster-RCNN和Level-Set结合的高分遥感影像建筑物提取

目前Level-Set图像分割方法存在初始轮廓的确定受人为因素影响较大的问题，对目标被遮盖和目标与背景灰度值相近无法达到理想的分割效果。针对此问题，本文提出了利用Faster-RCNN网络模型确定目标初始轮廓和区域信息的先验水平集图像分割方法，搭建Caffe深度学习框架训练Faster-RCNN网络模型；通过有监督学习的方式在IAILD数据集上训练模型，检测出目标建筑物并初步提取建筑物的轮廓，并将其与形状先验的Level-Set算法结合。对比实验结果表明，本文方法解决了Level-Set算法中图像分割结果初始轮廓受人为标记框选的影响较大的问题，能够更好地完成被遮挡建筑物的分割，对于目标建筑和背景灰度值相近也能达到更好的分割效果。     

融合先验形状知识的主动轮廓模型研究

融合先验形状知识的主动轮廓模型（HACM）能够解决目标存在遮挡、局部形变、相似变换等实际问题,广泛应用于图像分割、轮廓提取等领域。本文以主动轮廓模型（ACM）为基础,从先验形状知识的提取以及知识的融合方法两个方面,总结近年来ACM在融合先验形状知识的目标轮廓提取领域的一些最新研究成果。首先研究基于区域不变矩描述子、样条函数以及水平集的形状知识提取方法,用于提取刚体性、非刚体性目标样本集所包含的轮廓信息,该信息具有一定的泛化能力,即针对样本集以外的目标轮廓具备一定的表达能力。然后以单目标轮廓提取问题为研究对象,从距离函数的角度分析各种先验形状知识与原始轮廓模型的全局性融合方法,并介绍了一种基于标识函数的局部融合方法。局部融合方法可以解决单知识、多目标情况下的轮廓提取问题。最后对文中所介绍模型进行总结分析,对该领域的研究方向进行了展望。      

基于显著区域提取的红外图像舰船目标检测

针对复杂海面背景下红外图像舰船目标由于灰度不均匀、海杂波干扰大等因素造成的自动检测虚警率高、准确率低的问题,提出了一种显著区域提取和目标精确分割相结合的红外舰船目标检测方法。首先,利用基于图论的视觉显著性(Graph-based Visual Saliency ,GBVS)模型计算待检测图像的显著图,使得目标区域信息增强;其次,结合舰船目标先验信息（长短轴、面积等）,利用多级阈值划分算法提取关注的显著区域,并确定原图中候选目标区域;最后,利用空间约束模糊C均值(Fuzzy C-Means,FCM)算法对候选区域进行分割,结合目标先验知识对分割区域筛选并输出目标位置。所提方法在公开数据集IRShips上与相关方法进行比较,结果表明,相比直接进行全图目标搜索的方法,所提方法不仅准确率高、执行速度快,且检测目标的位置更加精确。     

基于形状角与分块图像的车辆轮廓分割算法

为了减小对车辆进行外形超限测量时其他车辆对图像的影响,精确地将目标车辆的轮廓信息分割并测量,提出一种基于形状角、结合分块图像与边缘提取的分割算法。对获取图像进行帧差法削弱背景边缘信息的影响,再用Canny算子边缘提取,将得到的轮廓信息进行基于形状角的识别与提取。提出了一种对图像进行分块分类的图像分割法。该算法有效解决了测量中背景对前景的影响。实验结果表明该算法具有较高的鲁棒性与准确性。     

基于稀疏表示的SAR地面目标分割技术

提出了一种引入先验约束的合成孔径雷达(SAR)图像的目标分割技术,以解决强杂波背景干扰下的目标分割困难问题。不同于基于统计理论的目标检测,文中利用目标图像切片在图像域的稀疏性,通过稀疏分解的方法构建目标特征窗函数实现目标的检测,并引入目标的形状先验对目标区域进行修正;然后,利用目标阴影的空间约束对基于统计检测的阴影区域进行修正,实现目标的分割;最后,基于实测数据验证了算法的有效性。     

基于交叉视觉皮质模型的非结构化道路检测算法

针对传统非结构化道路检测算法对各种复杂路面环境通用性不强且计算复杂度高的问题,提出了一种基于交叉视觉皮质模型(ICM)的道路检测算法.ICM具有接近生物视觉信息处理机制的特点,能够根据像素及其邻域的相关性动态区分目标和背景.基于ICM分割算法需要解决的问题是最佳分割阈值和循环迭代次数的确定,提出了采用最小交叉熵判决机制...     

Copolymer Networks Based on Poly(ω‐pentadecalactone) and Poly(ϵ‐caprolactone)Segments as a Versatile Triple‐Shape Polymer System

Thermo‐sensitive triple‐shape polymers can perform two consecutive shape changes in response to heat. These shape changes correspond to the recovery of two different deformations in reverse order, which were programmed previously at elevated temperature levels (T_mid and T_high) by the application of external stress. Recently, an AB copolymer network was described, which surprisingly exhibited a triple‐shape effect despite being programmed with only one deformation at T_high. Here it is explored whether a copolymer network system can be designed that enables a one‐step deformation process at ambient temperature (cold drawing) as a novel, gentle, and easy‐to‐handle triple‐shape‐creation procedure, in addition to the procedures reported to date, which generally involve deformation(s) at elevated temperature(s). A copolymer‐network system with two crystallizable polyester segments is synthesized and characterized, fulfilling two crucial criteria. These materials can be deformed at ambient temperature by cold drawing and show, even at T_high, which is above the melting points of both switching domains, elongation at break of up to 250%. Copolymer networks with PCL contents of 75 and 50 wt% show a triple‐shape effect after cold drawing with shape‐fixity ratios between 65% and 80% and a total‐shape‐recovery ratio above 97%. Furthermore, in these copolymer networks, the triple‐shape effect can be obtained after a one‐step deformation at T_high. Independent of the temperature at which the deformation is applied (ambient temperature or T_high), copolymer networks that have the same compositions show similar switching temperatures and proportioning of the recovery in two steps. The two‐step programming procedure enables a triple‐shape effect in copolymer networks for an even broader range of compositions. This versatile triple‐shape‐material system based on tailored building blocks is an interesting candidate material for applications in fixation systems or disassembling systems.     

一种基于微分思想的图象变形算法研究与实现

图象变形是图象处理中的基本问题,不仅要求图象变形效果好,而且要求变形算法效率高.目前已有的变形算法,大多采用基于象素点的填充方法,这种方法有比较明显的缺点:(1)当变形的图象区域是一个非规则区域时,图象变形处理所花费的时间增加,即算法的时间复杂度加大,效率低;(2)在对象素点作映射变换时,可能出现变形后的图象区域中一个象素点对应变形前的图象区域中多个象素点,因此造成取舍困难,甚至造成取舍错误.为解决这一问题,本文提出了一种基于微分思想的图象变形算法,该算法先将复杂的变形区域划分为一系列子区域,再将每个子区域划分为多个小矩形,将对应的小矩形按照标准矩形填充算法进行快速填充,由此实现图象变形.实验结果表明,该算法实现简单,计算速度快,填充效果好,并已成功地应用于笔者开发的基于真实照片的人脸整形与美容图象处理系统中.     

电子封装器件BGA的实时全息干涉实验研究

９０年代以来,表面封装器件（ＳＭＤ）向高密度化、精密化、薄型化和高集成化方向快速地发展。在研究这些新型电子封装的应力—应变实验技术中,全息干涉测量方法因其高灵敏度、高精度、非接触性、全场分析,可直接获取离面位移等优点而被作为表面安装技术（ＳＭＴ）可靠性分析的有效方法。本文采用实时全息干涉方法对加电运行中的ＰＢＧＡ器件及外部热辐射条件下ＰＢＧＡ的离面变形进行了测量,得到ＰＢＧＡ呈球面弯曲及马鞍形翘曲变形的实验结果,并初步分析了产生形变差异的原因。     

Polyampholyte Hydrogels with pH Modulated Shape Memory and Spontaneous Actuation

Polyampholyte hydrogels are synthesized by one‐step copolymerization of cationic monomer 3‐(methacryloylamino)propyltrimethylammonium chloride, anionic monomers sodium p‐styrenesulfonate (NaSS), and methacrylic acid (MAA) without chemical crosslinker and adding salts. The hydrogels exhibit pH responsive shape memory behavior; the temporary shape of the hydrogel is formed manually after immersing in NaOH solution and fixed in HCl solution, while the shape recovery occurs by immersing in NaOH again. Most interestingly, the hydrogel shows a spontaneous shape change after the first shape memory cycle. When the recovered hydrogel with a little residual deformation is immersed in HCl again, it twists spontaneously and rapidly to the previous temporary shape. The spontaneous twisting and recovering can be repeated for ten times. Furthermore, the hydrogel swells quickly and is strengthened in HCl, while shrinks and weakens in NaOH during the shape change procedure. This unique synergistic effect of fast swelling, residual helical deformation, and increased strength plays a significant role in the spontaneous shape alternation. This new finding will initiate a new prosperous design for new soft actuators requiring successive actions.     

Strain‐Based Temperature Memory Effect for Nafion and Its Molecular Origins

A shape memory polymer traditionally refers to a polymer that can memorize one temporary shape and recover to its permanent shape upon exposure to an external stimulus. Although this basic concept has been known for at least half a century, recent advances have led to the discoveriy of previously uncovered memory properties that challenge the traditional concept of shape memory polymers. In particular, a temperature memory effect refers to the capability of a polymer to memorize temperatures instead of shapes. Thus far, the reported temperature memory effect has been established under iso‐strain stress recovery conditions, in which the maximum recovery stress appears at a temperature roughly identical to the deformation temperature. This effect can be called recovery stress based temperature memory effect. In this work, experiments were designed in an attempt to establish a temperature memory effect based on the stress free strain recovery behaviors of Nafion. The results show that, under carefully selected conditions, the temperature at which a maximum strain recovery rate is observed can indeed be quantitatively related to the deformation temperature. In addition, indications that the polymer is capable of memorizing more than one deformation temperature (i.e., multi‐temperature memory effect) are shown. The molecular origin of Nafion’s temperature memory effect is elucidated through small angle neutron scattering study.     

螺旋布设光纤光栅的软体操作器形状传感方法

为实现微创手术软体操作器弯曲伸缩运动时的形状测量,提出一种螺旋布设光纤光栅形状传感方法。不同于传统的直线形布设方法,螺旋形布设方法可以防止软体操作器弯曲伸缩运动时,光纤光栅的折断、脱胶现象发生。理论分析了螺旋布设光纤光栅的传感原理及弯曲变形的重构算法;制备了光纤光栅沿螺旋形布设在软体操作器表面的试样;搭建了软体操作器弯曲变形的实验系统;实验分析了软体操作器在不同弯曲状态下,螺旋布设光纤光栅中心波长漂移量与弯曲角度间的变化规律,并重构出软体操作器的变形形状。实验结果表明:螺旋布设光纤光栅在软体操作器中不同位置的灵敏度最大为7.1pm/(°),软体操作器弯曲角度实际测量值与理论重构值间的最大误差不足4.29%。螺旋形光纤传感方法可用于软体操作器形状传感与重构。     

基于图像序列的非刚性物体运动形变程度估计方法

形变程度的定量分析是非刚性物体运动研究中一个重要问题.现有的形变程度估计方法一般都假设特征点位置误差是各向同性,并且不考虑特征点数据丢失的问题.通常事实并非如此,所以本文提出一种同时考虑图像的方向性误差和特征点数据丢失问题的形变程度估计方法.理论研究和人活动图像序列的实验表明,本文提出方法十分有效.     

Optical 3D shape measurement for dynamic process

3D shape dynamic measurement is essential to the study of machine vision, hydromechanics, high-speed rotation, deformation of material, stress analysis, deformation in impact, explosion process and biomedicine, in recent years. In this paper, the results of our research, including the theoretical analysis, some feasible methods and relevant verifying experiment results, are compendiously reported. At present, these results have been used in our assembling instruments for 3D shape measurement of dynamic process.     

Methods of artificial enlargement of the training set for statistical shape models

Due to the small size of training sets, statistical shape models often over-constrain the deformation in medical image segmentation. Hence, artificial enlargement of the training set has been proposed as a solution for the problem to increase the flexibility of the models. In this paper, different methods were evaluated to artificially enlarge a training set. Furthermore, the objectives were to study the effects of the size of the training set, to estimate the optimal number of deformation modes, to study the effects of different error sources, and to compare different deformation methods. The study was performed for a cardiac shape model consisting of ventricles, atria, and epicardium, and built from magnetic resonance (MR) volume images of 25 subjects. Both shape modeling and image segmentation accuracies were studied. The objectives were reached by utilizing different training sets and datasets, and two deformation methods. The evaluation proved that artificial enlargement of the training set improves both the modeling and segmentation accuracy. All but one enlargement techniques gave statistically significantly $(p ≪ 0.05)$ better segmentation results than the standard method without enlargement. The two best enlargement techniques were the nonrigid movement technique and the technique that combines principal component analysis (PCA) and finite element model (FEM). The optimal number of deformation modes was found to be near 100 modes in our application. The active shape model segmentation gave better segmentation accuracy than the one based on the simulated annealing optimization of the model weights.