多目标跟踪中目标数量准确识别算法

由于场景中存在目标遮挡与重叠的影响,以及检测时产生的分割错误,同一个目标往往被识别为若干运动区域,或者多个目标被识别为同一个运动区域,从而导致目标数量识别错误。为了解决这一问题,提出了一种结合融合与分离操作的理论来分析视频的方法,将多目标跟踪的问题转化成为一个找寻后验极值的问题。采用图论的方法表示观测结果,利用图像序列中目标运动轨迹和外观相似度的信息,将多目标跟踪的问题归结为找寻图中多个最优路径的问题。该方法采用了滑动窗口框架以便统计固定数目帧的信息。实验结果表明该方法能够应对实际中发生的上述现象,达到了准确识别多目标数量的目的。     

Redundant actuators to achieve minimal vibration trajectory tracking of flexible multibodies: Theory and application

We address the problem of inverse dynamics for flexible multibodies, which arises, in trajectory tracking control of flexible multibodies such as space manipulators and articulated flexible structures. Previous research has resolved this trajectory tracking problem by computing the system inputs for feedforward control of actuators at the joints. Recently, the use of distributed actuators like electro-strictive actuators in flexible structures has introduced a new dimension to this trajectory tracking problem. In this paper we optimally utilize such actuators to aid joint actuators for tracking control, and introduce a new inverse dynamics scheme for simultaneously (1) tracking a prescribed trajectory and (2) minimizing ensuing elastic deflections. We apply this scheme for trajectory tracking of a two-link two-joint planar manipulator with joint motors and distributed electro-strictive actuators. Experimental results are presented to contrast our new scheme with other existing methods.     

A Barban-Davenport-Halberstam theorem for integers with fixed number of prime divisors

The purpose of this paper is to study the distribution of integers with a given number prime divisors over arithmetic progressions, via using the large-sieve inequality, Huxley-Hooley contour and the zero-density estimate, and present a Barban-Davenport-Halberstam type theorem for it.     

Boundary contour analysis for surface stress recovery in 2-D elasticity and Stokes flow

Summary A variant of the boundary element method, called the boundary contour method, offers a further reduction in dimensionality. Consequently, boundary contour analysis of 2-D problems does not require any numerical integration at all. In a boundary contour analysis, boundary stresses can be accurately computed using the approach proposed in Ref. [1]. However, due to singularity, this approach can be used only to calculate boundary stresses at points that do not lie at an end of a boundary element. Herein, it is shown that a technique based on the displacement/velocity shape functions can overcome this drawback. Further, the approach is much simpler to apply, requires less computational effort, and provides competitive accuracy. Numerical solutions and convergence study for some well-known problems in linear elasticity and Stokes flow are presented to show the effectiveness of the proposed approach. This research was supported in part by the 2004 Ralph E. Powe Junior Faculty Enhancement Award from Oak Ridge Associated Universities and by the University of South Alabama Research Council.     

ELASTIC CALCULATION OF STRESSES IN RINGS USING AIRY STRESS FUNCTION

Stress calculation formulae for a ring have been obtained by using Airy stress function of the plane strain field with the decomposition of the solutions for normal stresses of Airy biharmonic equation into two parts when it is loaded under two opposite inside forces along a diameter. One part should fulfill a constraint condition about normal stress distribution along the circumference at an energy valley to do the minimum work. Other part is a stress residue constant. In order to verify these formulae and the computed results, the computed contour lines of equi-maximal shear stresses were plotted and quite compared with that of photo-elasticity test results. This constraint condition about normal stress distribution along circumference is confirmed by using Greens’ theorem. An additional compression exists along the circumference of the loaded ring, explaining the divorcement and displacement of singularity points at inner and outer boundaries.     

Three-dimensional numerical simulation of drops suspended in simple shear flow at finite Reynolds numbers

A three-dimensional study of suspension of drops in simple shear flow has been performed at finite Reynolds numbers. Results are obtained using a finite difference/front tracking method in a periodic domain. The effects of the Reynolds number and the Capillary number are addressed at two volume fractions: 0.195 and 0.34. It is observed that suspensions of deformable drops exhibit a shear-thinning behavior. Similar to the motion of a single drop, drops migrate away from the walls. The effective viscosity, the first and the second normal stress differences oscillate around a mean value in all cases. The first normal stress difference increases with the Capillary number, the Reynolds number and the volume fraction. Results show that drops deform more and orient more in the flow direction as the Capillary number or the volume fraction is increased. Also, the average size of clusters is smaller than for suspension of rigid particles. The radial dependence of the pair distribution function across the channel has been studied. This dependency shows that the tendency to form clusters is reduced as the Capillary number increases or the volume fraction decreases.     

改进的粒子滤波算法在目标跟踪中的应用

粒子滤波算法中通常采用先验转移概率代替重要性函数,由此重要性密度函数对后验函数的偏差将增大。将小波去噪应用到粒子滤波过程中,降低了偏差,提高了粒子算法的滤波精度,并将该算法应用到目标跟踪的过程中,通过仿真证实该方法能够提高粒子滤波精度。     

一种结合Mean—shift和粒子滤波的视频跟踪算法

提出一种结合均值偏移算法和粒子滤波理论的视频跟踪算法,解决了目标旋转、部分遮挡和运动速度过快的问题．通过均值偏移对粒子滤波中的粒子集进行进一步寻优,增加了粒子的有效性,极大减少了粒子采样的数量,且解决了经过多次重采样后粒子多样性降低的问题．新的粒子通过与观测值之间的巴氏系数来决定重要性权重．实验证明：本算法可以完成实时地对视频目标进行部分遮挡以及目标旋转下的跟踪,具有较强的鲁棒性．     

Mathematical modelling of bottom evolution by particle deposition and resuspension

A two-dimensional numerical model for the evolution of a bottom due to particle deposition and resuspension by a fluid flow is here presented. A computational fluid dynamic approach is used to calculate the flow field and a Lagrangian particle tracking technique is applied to solve the dispersed phase. The evolution of the lower boundary is simulated taking into account the mass conservation of the solid phase and the geotechnical properties of the granular material. The model is characterized by two important features. First, fluid dynamics are coupled with the bottom evolution due to particle deposition and resuspension. This permits to use the model to simulate complex flow fields as well as complex time-evolving geometries. Second, the dispersed phase is calculated by a Lagrangian approach, which retains the discrete information of the individual particles of the granular bottom which may be of interest for some industrial processes (coating) and environmental flows (sediment stratification). First consistency checks have been performed for some deposition and resuspension test cases with fluid at rest. The model has also been tested by comparison with a physical experiment of deposition inside a cavity. Finally, as an example of possible applications of industrial and environmental interest, the model has been applied to investigate particle deposition in rectangular cavities and the evolution of a sand heap by a fluid flow.     

A correlative classifiers approach based on particle filter and sample set for tracking occluded target

Target tracking is one of the most important issues in computer vision and has been applied in many fields of science, engineering and industry. Because of the occlusion during tracking, typical approaches with single classifier learn much of occluding background information which results in the decrease of tracking performance, and eventually lead to the failure of the tracking algorithm. This paper presents a new correlative classifiers approach to address the above problem. Our idea is to derive a group of correlative classifiers based on sample set method. Then we propose strategy to establish the classifiers and to query the suitable classifiers for the next frame tracking. In order to deal with nonlinear problem, particle filter is adopted and integrated with sample set method. For choosing the target from candidate particles, we define a similarity measurement between particles and sample set. The proposed sample set method includes the following steps. First, we cropped positive samples set around the target and negative samples set far away from the target. Second, we extracted average Haar-like feature from these samples and calculate their statistical characteristic which represents the target model. Third, we define the similarity measurement based on the statistical characteristic of these two sets to judge the similarity between candidate particles and target model. Finally, we choose the largest similarity score particle as the target in the new frame. A number of experiments show the robustness and efficiency of the proposed approach when compared with other state-of-the-art trackers.