Scene-adaptive hierarchical data association and depth-invariant part-based appearance model for indoor multiple objects tracking

Indoor multi-tracking is more challenging compared with outdoor tasks due to frequent occlusion, view-truncation, severe scale change and pose variation, which may bring considerable unreliability and ambiguity to target representation and data association. So discriminative and reliable target representation is vital for accurate data association in multi-tracking. Pervious works always combine bunch of features to increase the discriminative power, but this is prone to error accumulation and unnecessary computational cost, which may increase ambiguity on the contrary. Moreover, reliability of a same feature in different scenes may vary a lot, especially for currently widespread network cameras, which are settled in various and complex indoor scenes, previous fixed feature selection schemes cannot meet general requirements. To properly handle these problems, first, we propose a scene-adaptive hierarchical data association scheme, which adaptively selects features with higher reliability on target representation in the applied scene, and gradually combines features to the minimum requirement of discriminating ambiguous targets; second, a novel depth-invariant part-based appearance model using RGB-D data is proposed which makes the appearance model robust to scale change, partial occlusion and view-truncation. The introduce of RGB-D data increases the diversity of features, which provides more types of features for feature selection in data association and enhances the final multi-tracking performance. We validate our method from several aspects including scene-adaptive feature selection scheme, hierarchical data association scheme and RGB-D based appearance modeling scheme in various indoor scenes, which demonstrates its effectiveness and efficiency on improving multi-tracking performances in various indoor scenes.

Effective method for tracking multiple objects in real-time visual surveillance systems

An object model-based tracking method is useful for tracking multiple objects, but the main difficulties are modeling objects reliably and tracking objects via models in successive frames. An effective tracking method using the object models is proposed to track multiple objects in a real-time visual surveillance system. Firstly, for detecting objects, an adaptive kernel density estimation method is utilized, which uses an adaptive bandwidth and features combining colour and gradient. Secondly, some models of objects are built for describing motion, shape and colour features. Then, a matching matrix is formed to analyze tracking situations. If objects are tracked under occlusions, the optimal ＂visual＂ object is found to represent the occluded object, and the posterior probability of pixel is used to determine which pixel is utilized for updating object models. Extensive experiments show that this method improves the accuracy and validity of tracking objects even under occlusions and is used in real-time visual surveillance systems.

Group Decision-Making Method with Incomplete Intuitionistic Fuzzy Preference Relations Based on a Generalized Multiplicative Consistent Concept

Based on the analyses of existing preference group decision-making(PGDM)methods with intuitionistic fuzzy preference relations(IFPRs),we present a new PGDM framework with incomplete IFPRs.A generalized multiplicative consistent for IFPRs is defined,and a mathematical programming model is constructed to supplement the missing values in incomplete IFPRs.Moreover,in this study,another mathematical programming model is constructed to improve the consistency level of unacceptably multiplicative consistent IFPRs.For group decisionmaking(GDM)with incomplete IFPRs,three reliable sources influencing the weights of experts are identified.Subsequently,a method for determining the weights of experts is developed by simultaneously considering three reliable sources.Furthermore,a targeted consensus process(CPR)is developed in this study with reference to the actual situation of the consensus level of each IFPR.Meanwhile,in response to the proposed multiplicative consistency definition,a novel method for determining the optimal priority weights of alternatives is redefined.Lastly,based on the above theory,a novel GDM method with incomplete IFPRs is developed,and the comparative and sensitivity analysis results demonstrate the utility and superiority of this work.

A trajectory data warehouse solution for workforce management decision-making

In modern workforce management,the demand for new ways to maximize worker satisfaction,productivity,and security levels is endless.Workforce movement data such as those source data from an access control system can support this ongoing process with subsequent analysis.In this study,a solution to attaining this goal is proposed,based on the design and implementation of a data mart as part of a dimensional trajectory data warehouse(TDW)that acts as a repository for the management of movement data.A novel methodological approach is proposed for modeling multiple spatial and temporal dimensions in a logical model.The case study presented in this paper for modeling and analyzing workforce movement data is to support human resource management decision-making and the following discussion provides a representative example of the contribution of a TDW in the process of information management and decision support systems.The entire process of exporting,cleaning,consolidating,and transforming data is implemented to achieve an appropriate format for final import.Structured query language(SQL)queries demonstrate the convenience of dimensional design for data analysis,and valuable information can be extracted from the movements of employees on company premises to manage the workforce efficiently and effectively.Visual analytics through data visualization support the analysis and facilitate decisionmaking and business intelligence.

用SuperMap Object开发基于GIS的连锁店配送信息系统

支撑整个连锁店运营过程的是基于GIS的连锁店配送信息系统。在明确了业务流程的基础上,研究了基于GIS的连锁店配送信息系统的功能、GIS平台、空间数据模型、道路数据模型和最优路径分析,探讨了以SuperMapObjects作为GIS平台,采用MicrosoftVisualBasic作为开发工具实现基于GIS的连锁店配送信息系统的方案。

An ideal interval method of multi-objective decision-making for comprehensive assessment of water resources renewability

In order to estimate water resources renewability scientifically, an Ideal Interval Method of Multiple Objective Decision-Making (IIMMODM) is presented. This method is developed through improving an ideal point method of multiple objective decision-making. The ideal interval is obtained with assessment standard instead of ideal points. The weights are decided by using the basic point and gray code accelerating genetic algorithm. This method has synthesized the expert’s suggestion and avoided giving a mark for the objective again. It could solve the complicated problem of compatible or incompatible multi-objective assessment. The principle of IIMMODM is presented in this paper. It is used to assess the water resources renewability for nine administrative divisions in the Yellow River basin. The result shows that the water resources renewability in the Yellow River basin is very low. Compared with the gray associate analysis method, fuzzy synthesis method and genetic projection pursuit method, the IIMMODM is easier to use. Compared with the ideal point method of multiple objective decision-making, the IIMMODM has good robustness, which is applicable to the comprehensive assessments of water resources.

森林生态系统服务多功能性:概念、指标和经营模拟模型

森林生态系统同时提供多种功能和服务的能力,被称为森林生态系统多功能性,可分为生态系统功能多功能性和生态系统服务多功能性。近年来,生态学领域对生物多样性与多功能性关系机理、多功能性驱动因子等进行了大量研究,这些研究持续成为热点。然而,这些研究成果在实际应用,尤其是在指导森林经营实践方面,尚显不足。森林经营一直以实现多种功能或服务的最大化为目标,但目前还缺乏与生态学中多功能性理论研究的有机联系,迫切需要将二者结合。本文对国际国内森林生态系统服务多功能性的概念、指标、评价方法、经营模拟和优化模型进行了综述,认为预测森林生态系统服务及多功能性、在森林经营单位层面实现多功能性最大化是将来的研究方向。因此,需要加强森林经营模型模拟研究,寻找减少不同服务间冲突、增加协同(即增加多功能性)的经营策略,回答森林经营和气候变化如何影响生态系统服务间的权衡和协调关系及其时空变化等理论和实际问题。

System reliability-based robust design of deep foundation pit considering multiple failure modes

Recently,reliability-based design is a universal method to quantify negative influence of uncertainty in geotechnical engineering.However,for deep foundation pit,evaluating the system safety of retaining structures and finding cost-effective design points are main challenges.To address this,this study proposes a novel system reliability-based robust design method for retaining system of deep foundation pit and illustrated this method via a simplified case history in Suzhou,China.The proposed method included two parts:system reliability model and robust design method.Back Propagation Neural Network(BPNN)is used to fit limit state functions and conduct efficient reliability analysis.The common source random variable(CSRV)model are used to evaluate correlation between failure modes and determine the system reliability.Furthermore,based on the system reliability model,a robust design method is developed.This method aims to find cost-effective design points.To solve this problem,the third generation non-dominated genetic algorithm(NSGA-III)is adopted.The efficiency and accuracy of whole computations are improved by involving BPNN models and NSGA-III algorithm.The proposed method has a good performance in locating the balanced design point between safety and construction cost.Moreover,the proposed method can provide design points with reasonable stiffness distribution.

一种基于IMM的分布式扩展目标跟踪算法

随着传感器分辨率的提高,将目标视为一个质点可能导致大量信息的丢失,传统点目标跟踪模型不再适用。而扩展目标跟踪算法不仅考虑了目标的运动状态(如位置、速度和加速度),还考虑了目标的扩展状态(如形状、大小和方向等),可获得更准确且完整的目标状态估计。近年来,利用随机矩阵跟踪扩展目标的方法颇受欢迎。实际场景中扩展目标运动复杂多变,可能导致其运动状态和扩展发生突变。多模型算法,如交互多模型(Interactive Multiple Model,IMM)算法,是一种有效的机动目标跟踪方法。本文考虑利用传感器网络以分布式算法实现机动扩展目标的跟踪问题。本文提出了一种基于扩散策略的分布式机动扩展目标跟踪算法,其中,采用随机矩阵法对扩展目标进行建模。该算法拓展了IMM框架,以描述不同扩展特性的扩展目标在不同过程噪声下的运动特性,并进一步研究了一种减小通信量的分布式局部扩散策略。具体地说,在该算法中,每个节点基于IMM框架跟踪机动扩展目标,并采用加权Kullback-Leibler平均实现IMM框架中的数据融合。此外,应用本文所提出的局部扩散策略,每个节点仅与邻居节点交换部分中间估计值,以实现较低通信负担的有效的分布式扩展目标跟踪。仿真实验结果表明,本文所提的基于局部扩散策略的分布式机动扩展目标跟踪算法能够有效地跟踪机动扩展目标,且具有相对较低的网络通信负担。

高速公路ETC门架和收费站车辆轨迹数据数学模型研究

采用蒙特卡洛法对高速公路车辆轨迹数据进行预处理,模拟得出车流量的时间分布特征和空间分布特征.采用二分法遍历车辆轨迹数据,得出各电子收费系统(ETC)门架的漏失记录率以及最易被漏失的车型.建立收费站车流量与所有收费站总车流量的多元线性回归模型,得出收费站车流量与所有收费站总车流量的关系.建立各收费站车道数量和应急收费系统数量的单目标规划模型,得出收费站通行顺畅且车道和应急收费系统建设费用最少时,各收费站车道数量以及应急收费系统数量.采用G(1,1)灰色模型预测出15个收费站间各路段未来24 h的车流量变化趋势,其与实际车流量变化趋势基本吻合,因此可将各路段预测车流量作为参考,为高速公路管理部门改进交通状况提供改进依据.

概念性水文模型多目标参数自动优选方法研究

以三水源新安江模型为例,研究探讨了概念性水文模型多目标参数自动优选方法。目标函数综合考虑了水量平衡、确定性系数、洪峰和枯水流量过程。通过对目标函数的不同组合方式,分别率定模型参数和分析比较结果。研究表明,多目标参数自动优选方法综合考虑了水文过程的各种要素,优于传统的单目标优选结果,具有较高的模拟预报精度。

基于同类城市建设目标的北京市土地利用结构优化

基于北京与东京土地利用的对比分析,对北京市土地利用结构进行优化,更好地发挥城市土地的宏观调控功能。根据北京市1992-2012年和东京1982-2011年多期土地利用现状数据,集成多目标优化模型的目标决策功能和元胞自动机-马尔可夫链(CA-Markov)模型的空间布局特点,对北京市2020、2030年的土地利用进行数量结构优化和空间布局模拟。通过对北京与东京在土地利用演变过程的深入分析,归纳总结出北京土地利用的发展定位以及2020、2030年的定量目标,然后对土地利用经济、社会和生态3方面构建多目标模型,最后对模型进行求解,同时应用CA-Markov模型对北京市2020、2030年优化后的土地利用结构进行空间布局模拟,通过运算,北京市农用地、生态用地、建设用地、交通用地、水域、其他土地在2020年、2030年的比例分别为21.00%、50.50%、19.00%、3.00%、4.50%、2.00%和20.00%、50.00%、20.00%、3.50%、4.50%、2.00%。优化后的土地利用结构改变以往剧烈变化的现象,更趋向于稳定状态,优化后的土地利用布局更为集约并兼顾生态保护功能。

一种地区电网多目标无功优化的新方法——改进模拟退火算法

本文介绍了改进模拟退火算法在地区电网无功优化中的应用。实际计算表明，与常规优化方法比较，模拟退火算法收敛性好、适应性强，是实现无功优化的好方法。另外，本文给出了一种新的多目标优化模型，测试结果验证了此法的可行性。

一个改进的可动态调节的机密性策略模型

试图提出一个模型,它能为有效处理网络安全对象提供支持.改进Amon ott的动态调整规则,使需要动态改变的量减少,从而使一个系统调用级的原子操作需要伴随的附加量的操作和存储减少,提高模型在系统中的实用性.通过把当前安全级变成敏感标签范围而增加模型在系统实现中的灵活性,能有效控制IPC对象.为此,把Amon ott动态地调整当前敏感标签的规则推广为动态地调整敏感标签范围的规则,这把Bell为处理网络情况而把主体的当前安全级变成敏感标签范围的工作与Amon ott的工作有机地结合起来,同时,参考实际中使用的系统GEMSOS和DG/UX及安全原型微内核系统Fluke,引入了单级实体、多级实体以及专用于进程的访问模式,并给出了它们应满足的不变量及限制性条件.另外,在参考原型系统TMach对IPC对象的某些处理方式的基础上,提出了使动态特征非常明显的IPC对象得到合理、有效管理的机制,同时还分析了ABLP实施方法中存在的一些不当之处.提出了一个新的机密性策略模型DBLP的模型不变量,限制性条件及变量类型和常量,并分析了一些限制性条件的合理性,它可以有效应用于系统设计.

活动轮廓模型目标跟踪算法综述

目标跟踪是当前计算机视觉领域最活跃的研究主题。首先对基本的跟踪类型进行了介绍;然后讨论了基于活动轮廓模型的图像分割,重点分析了参数活动轮廓模型的梯度矢量流模型(Gradient Vector Flow,GVF),以及几何活动轮廓模型中的模型;并讨论了基于粒子滤波的目标跟踪算法的研究现状,最后展望了这一领域未来研究的热点。

多属性决策中权重确定的主客观赋权法

研究了多属性决策中属性指标权重问题的确定，提出了一种主客观赋权法，该方法是通过一个数学规划模型自动求解属性的权重，使确定的权重同时反映主观程度和客观程度．

多目标混流装配计划排序问题

在总结混流装配线排序问题的优化目标基础上,给出了一种混流装配线排序多目标优化模型.采用一种染色体映射的策略,有效地解决了遗传算法中染色体变化后难以保证主生产计划约束的问题.对比计算的结果表明:多目标计划排序不仅能有效保证基于生产负荷平衡的目标和基于物流平顺化的目标的要求,同时还能减少装配中品种切换频繁的问题.

基于快速FCM算法的多目标分割CV模型

Chan-Vese(CV)模型是基于水平集方法演化不依赖图像梯度的算法,能很好地处理拓扑变化和弱边界,但对于目标和背景对比度低的边界以及多目标区域分割效果较差。针对上述问题提出一种基于快速模糊F均值(FCM)算法和邻域模板改进的CV模型。利用快速FCM算法提取图像特征信息,采用邻域模板阈值法对不同的目标区域分别处理,准确控制了轮廓线的分裂,能够分割出更多的目标区域。

基于属性关系图优化匹配的多运动目标跟踪

针对复杂监控场景中往往无法准确检测前景目标,导致难以有效跟踪目标的问题,提出了一种基于属性关系图优化匹配的多运动目标跟踪方法,将目标跟踪问题转化为前景目标跟踪标记的优化问题,实现多运动目标跟踪。将前景区域按颜色、空间特征一致性划分为多个碎片;采用属性关系图描述目标模型,分析计算出属性关系图外观模型的属性相似度;提出通过概率松弛法计算目标函数并采用遗传算法进行优化匹配,得到前景碎片最优的跟踪标记,从而完成多目标跟踪。在多个监控视频上的实验结果表明,本方法能大大提高跟踪性能,实现复杂监控场景中的多目标跟踪。

考虑双边主体公平性的稳定匹配决策方法

针对一对一双边匹配决策问题,依据双边主体给出的多指标评价信息提出了一种新的双边匹配决策方法。给出了稳定匹配和公平匹配的相关概念,并给出了双边主体满意度的计算方法;建立了考虑稳定性、公平性和满意性的多目标优化模型,并设计了一个多目标遗传算法用于求解模型。最后,通过教育服务机构中学生与老师的双边匹配实例验证了方法的实用性和有效性。