Matching actions in presence of camera motion

When the camera viewing an action is moving, the motion observed in the video not only contains the motion of the actor but also the motion of the camera. At each time instant, in addition to the camera motion, a different view of the action is observed. In this paper, we propose a novel method to perform action recognition in presence of camera motion. Proposed method is based on the epipolar geometry between any two views. However, instead of relating two static views using the standard fundamental matrix, we model the motions of independently moving cameras in the equations governing the epipolar geometry and derive a new relation which is referred to as the “temporal fundamental matrix.” Using the temporal fundamental matrix, a matching score between two actions is computed by evaluating the quality of the recovered geometry. We demonstrate the versatility of the proposed approach for action recognition in a number of challenging sequences.     

Geometry and construction of straight lines in log-polar images

This paper investigates, how straight lines are mapped to the non-Euclidean log-polar image plane and how their properties in log-polar coordinates can be used to perform two fundamental measurement operations: the construction of straight lines of given orientation angle and the measurement of distances along these lines. It is shown, how the two operations can be efficiently implemented for discrete log-polar images. The usefulness of the developed theory is demonstrated by applying it to different image processing tasks. As a first example, it is used for line and circle detection in log-polar sampled images. As a second example, it is used to construct epipolar lines and perform disparity estimation in log-polar images. Experiments with both synthetic and real images are presented, and the feature detection results are quantitatively evaluated.     

A hierarchy of cameras for 3D photography

The view-independent visualization of 3D scenes is most often based on rendering accurate 3D models or utilizes image-based rendering techniques. To compute the 3D structure of a scene from a moving vision sensor or to use image-based rendering approaches, we need to be able to estimate the motion of the sensor from the recorded image information with high accuracy, a problem that has been well-studied. In this work, we investigate the relationship between camera design and our ability to perform accurate 3D photography, by examining the influence of camera design on the estimation of the motion and structure of a scene from video data. By relating the differential structure of the time varying plenoptic function to different known and new camera designs, we can establish a hierarchy of cameras based upon the stability and complexity of the computations necessary to estimate structure and motion. At the low end of this hierarchy is the standard planar pinhole camera for which the structure from motion problem is non-linear and ill-posed. At the high end is a camera, which we call the full field of view polydioptric camera, for which the motion estimation problem can be solved independently of the depth of the scene which leads to fast and robust algorithms for 3D Photography. In between are multiple view cameras with a large field of view which we have built, as well as omni-directional sensors.     

A practical approach for active camera coordination based on a fusion-driven multi-agent system

In this paper, we propose a multi-agent system architecture to manage spatially distributed active (or pan-tilt-zoom) cameras. Traditional video surveillance algorithms are of no use for active cameras, and we have to look at different approaches. Such multi-sensor surveillance systems have to be designed to solve two related problems: data fusion and coordinated sensor-task management. Generally, architectures proposed for the coordinated operation of multiple cameras are based on the centralisation of management decisions at the fusion centre. However, the existence of intelligent sensors capable of decision making brings with it the possibility of conceiving alternative decentralised architectures. This problem is approached by means of a MAS, integrating data fusion as an integral part of the architecture for distributed coordination purposes. This paper presents the MAS architecture and system agents.     

Constructing invariants for hybrid systems

We present a new method for generating algebraic invariants of hybrid systems. The method reduces the invariant generation problem to a constraint solving problem using techniques from the theory of ideals over polynomial rings. Starting with a template invariant—a polynomial equality over the system variables with unknown coefficients—constraints are generated on the coefficients guaranteeing that the solutions are inductive invariants. To control the complexity of the constraint solving, several stronger conditions that imply inductiveness are proposed, thus allowing a trade-off between the complexity of the invariant generation process and the strength of the resulting invariants. This research was supported in part by NSF grants CCR-01-21403, CCR-02-20134 and CCR-02-09237, by ARO grant DAAD19-01-1-0723, by ARPA/AF contracts F33615-00-C-1693 and F33615-99-C-3014, and by NAVY/ONR contract N00014-03-1-0939.     

基于单张平行六面体照片的摄像机标定方法

摄像机标定是计算机视觉的重要组成部分。根据射影几何理论建立了基于单张平行六面体照片的摄像机全参数标定方法。通过调和射影及Laguerre定理的推论求解出平面圆环点的共轭虚像,利用平行六面体照片存在的三对圆环点的虚像,构造对无穷远平面绝对二次曲线的像ω的约束,使用辗转相除法求解非线性问题。最后,给出该算法对模拟图像和真实照片的实验。实验结果表明计算结果与真实情况吻合。     

视觉基础矩阵估计方法的性能比较与分析

基础矩阵描述了单个场景的2幅视图之间的对应关系,在计算机视觉领域中扮演着极其重要的角色.首先讨论了三维视觉系统中的极线几何理论,随后论述了几类基础矩阵的估计方法,并对这些方法作了详细的比较和评价,最后实现了各种算法且使用仿真数据以及真实图像数据对各自的性能作了分析和总结.比较结果说明:1)如果图像特征点定位精确并且匹配无误,则线性方法的结果相当好;2)迭代算法可以解决定位的高斯噪声,但是当数据存在错误匹配时,性能很差;3)鲁棒算法能够同时解决定位误差和误匹配两类问题.此外,模拟实验和真实图像实验的结果表明,基于特征分析的正交回归最小二乘法的计算结果优于经典的线性最小二乘法.     

Specifications for decidable hybrid games

We introduce STORMED hybrid games (SHG), a generalization of STORMED hybrid systems, Vladimerou et al. (2008) [33], which have natural specifications that allow rich continuous dynamics and various decidable properties. We solve the control problem for SHG using a reduction to bisimulation on finite game graphs. This generalizes to a greater family of games, which includes o-minimal hybrid games, Bouyer et al. (2006) [6]. We also solve the optimal-cost reachability problem for Weighted SHG and prove decidability of WCTL for Weighted STORMED hybrid systems.     

Abstractions for hybrid systems

We present a procedure for constructing sound finite-state discrete abstractions of hybrid systems. This procedure uses ideas from predicate abstraction to abstract the discrete dynamics and qualitative reasoning to abstract the continuous dynamics of the hybrid system. It relies on the ability to decide satisfiability of quantifier-free formulas in some theory rich enough to encode the hybrid system. We characterize the sets of predicates that can be used to create high quality abstractions and we present new approaches to discover such useful sets of predicates. Under certain assumptions, the abstraction procedure can be applied compositionally to abstract a hybrid system described as a composition of two hybrid automata. We show that the constructed abstractions are always sound, but are relatively complete only under certain assumptions.     

基于资源配置混杂Petri网的混杂系统生产过程建模

为了增强混杂Ｐｅｔｒｉ网解决资源共享和资源冲突的能力,定义一种新的混杂Ｐｅｔｒｉ网模型———资源配置混杂Ｐｅｔｒｉ网,提出了相应的使能和激发规则．将对连续变迁和离散变迁的控制作用引入混杂Ｐｅｔｒｉ网,同时,增加了资源配置变迁和资源释放变迁,用于有效分配可重复利用的资源．以典型的混杂生产过程为例,研究混杂系统生产过程建模．研究结果表明,所定义的模型描述能力强,模型语义正确合理,能够有效描述和分析混杂系统生产过程．     

Output-to-state stability for hybrid systems

Output-to-state stability (OSS) is a dual notion of input-to-state stability for dynamical systems. This paper presents Lyapunov and asymptotic characterizations of OSS for hybrid dynamical systems, emphasizing that a globally detectable (i.e. nonuniformly OSS) hybrid system admits a smooth OSS-Lyapunov function.     

Kalman filter for controlled hybrid systems

The linear partially observed discrete-continuous (hybrid) stochastic controllable system described by differential equations with measures is considered. The optimal filtering equations in the form of generalized Kalman filter are obtained in the case of non-anticipating control. This result could be a theoretical basis for the optimal control in stochastic hybrid systems with incomplete information.     

Model simplification for switched hybrid systems

This paper is concerned with the problem of model reduction for switched system, which is an important class of hybrid systems frequently encountered in practical situations. Two sharply different approaches are proposed to solve this problem. The first approach casts the model reduction into a convex optimization problem, which is the first attempt to solve the model reduction problem by using linearization procedure. The second one, based on the cone complementarity linearization idea, casts the model reduction problem into a sequential minimization problem subject to linear matrix inequality constraints. Both approaches have their own advantages and disadvantages concerning conservatism and computational complexity. A numerical example illustrates the effectiveness of the proposed theories.     

Automated synthesis of hybrid Petri net models for robotic cells in the aircraft industry

The use of robots in the aircraft structural assembly is a challenge. The presence of human operators, auxiliary systems and industrial robots makes hybrid the dynamic behavior of a robotic cell in this context. Here, the focus is on the automated synthesis of a model for the sequencing of the activities of a robotic cell in the aircraft industry. The cell model is obtained from a simple specification of resources and tasks, considered the main cell components, running the algorithm presented in this paper. The effectiveness of the model is shown using a case study defined by the ongoing European project LOCOMACHS (LOw COst Manufacturing and Assembly of Composite and Hybrid Structures, http://www.locomachs.eu/).     

A hybrid receding-horizon control scheme for nonlinear systems

A hybrid receding-horizon control scheme for nonlinear discrete-time systems is proposed. Whereas a set of optimal feedback control functions is defined at the continuous level, a discrete-event controller chooses the best control action, depending on the current conditions of a plant and on possible external events. Such a two-level scheme is embedded in the structure of abstract hybrid systems, thus making it possible to prove a new asymptotic stability result for the hybrid receding-horizon control approach.     

Safety verification and reachability analysis for hybrid systems

Safety verification and reachability analysis for hybrid systems is a very active research domain. Many approaches that seem quite different, have been proposed to solve this complex problem. This paper presents an overview of various approaches for autonomous, continuous-time hybrid systems and presents them with respect to basic problems related to verification.     

Gronwall inequality for hybrid systems

The Gronwall inequality, a well-known and useful result both for continuous-time and discrete-time signals, is extended to hybrid signals, namely those that combine continuous time and discrete time. An application of the result to establishing a bounded energy bounded state property for hybrid systems with inputs is provided.     

混杂系统的预测控制

因混杂系统预测控制（PC-HS）具有在线处理控制量和状态量的约束等优点,而成为学术界研究的一个热点．系统地论述了PC-HS的发展和研究现状,在介绍混杂系统的背景和模型的基础上,重点讨论了PC-HS的各种算法、性能等几个前沿问题,并对PC-HS目前研究中存在的问题进行了分析,对未来的一些研究方向进行了展望．     

A hybrid multiple feature construction approach for classification using Genetic Programming

The purpose of feature construction is to create new higher-level features from original ones. Genetic Programming (GP) was usually employed to perform feature construction tasks due to its flexible representation. Filter-based approach and wrapper-based approach are two commonly used feature construction approaches according to their different evaluation functions. In this paper, we propose a hybrid feature construction approach using genetic programming (Hybrid-GPFC) that combines filter’s fitness function and wrapper’s fitness function, and propose a multiple feature construction method that stores top excellent individuals during a single GP run. Experiments on ten datasets show that our proposed multiple feature construction method (Fcm) can achieve better (or equivalent) classification performance than the single feature construction method (Fcs), and our Hybrid-GPFC can obtain better classification performance than filter-based feature construction approaches (Filter-GPFC) and wrapper-based feature construction approaches (Wrapper-GPFC) in most cases. Further investigations on combinations of constructed features and original features show that constructed features augmented with original features do not improve the classification performance comparing with constructed features only. The comparisons with three state-of-art methods show that in majority of cases, our proposed hybrid multiple feature construction approach can achieve better classification performance.     

A computational approach to synthesizing guards for hybrid systems

We propose a technique for synthesizing switching guards for hybrid systems to satisfy a given state-based safety constraint. Using techniques from sum of squares (SOS) optimization, we design guards defined by semialgebraic sets that trigger mode switches, and we guarantee that the synthesized switching policy does not allow Zeno executions. We demonstrate our approach on an example of switched affine systems and on an application to traffic ramp metering.