An Efficient Particle Filter–based Tracking Method Using Graphics Processing Unit (GPU)

Particle filter has been proven very robust in handling non-linear and non-Gaussian problems and has been widely used in the area of object tracking. One of the main problems in particle filter-based object tracking is, however, its high computational cost induced by the most time-consuming stage of measurement model computation. This paper makes progress in resolving the problem by proposing an efficient particle filter-based tracking algorithm using color information. First, a compact color cooccurrence histogram is presented, which considers both spatial and color information and can effectively represent color distribution with a very small number of histogram bins. The paper also introduces integral images by which the cooccurrence histogram can be obtained with simple array reference operations. However, the construction of the integral images on the CPU may be computationally expensive. Hence, this paper develops parallel algorithms on a desktop Graphics Processing Unit (GPU), which accomplishes the integral images construction and cooccurrence histogram computation after bin index determination. The resulting algorithm is quite efficient and has better performance than the traditional histogram-based tracking algorithm. The tracking time of the proposed algorithm increases insignificantly with the growth of particle number, and it remains consistent among varying image sequences and stable throughout all frames in the same image sequence due to its irrelevance to object size. Experiments in diverse image sequences validate our conclusions.     

A Robust Particle Filter-Based Method for Tracking Single Visual Object Through Complex Scenes Using Dynamical Object Shape and Appearance Similarity

This paper addresses the issue of tracking a single visual object through crowded scenarios, where a target object may be intersected or partially occluded by other objects for a long duration, experience severe deformation and pose changes, and different motion speed in cluttered background. A robust visual object tracking scheme is proposed that exploits the dynamics of object shape and appearance similarity. The method uses a particle filter where a multi-mode anisotropic mean shift is embedded to improve the initial particles. Comparing with the conventional particle filter and mean shift-based tracking (Shan et al. 2004), our method offers the following novelties: We employ a fully tunable rectangular bounding box described by five parameters (2D central location, width, height, and orientation) and full functionaries in the joint tracking scheme; We derive the equations for the multi-mode version of the anisotropic mean shift where the rectangular bounding box is partitioned into concentric areas, allowing better tracking objects with multiple modes. The bounding box parameters are then computed by using eigen-decomposition of mean shift estimates and weighted averaging. This enables a more efficient re-distributions of initial particles towards locations associated with large weights, hence an efficient particle filter tracking using a very small number of particles (N = 15 is used). Experiments have been conducted on video containing a range of complex scenarios, where tracking results are further evaluated by using two objective criteria and compared with two existing tracking methods. Our results have shown that the propose method is robust in terms of tracking drift, tightness and accuracy of tracked bounding boxes, especially in scenarios where the target object contains long-term partial occlusions, intersections, severe deformation, pose changes, or cluttered background with similar color distributions.     

一种改进型的粒子滤波器

针对复杂背景条件下图像序列中跟踪运动目标的问题,提出一种改进的粒子滤波图像跟踪算法,该算法利用遗传算法的研究成果,采用选择、交叉、变异等步骤实现对粒子的重采样,解决了粒子滤波器所面临的粒子退化和匮乏问题.由于该算法利用了遗传算法的全局寻优特性,因此该算法具有较强的稳健性.同时,粒子滤波可实现非线性非高斯状态空间模型的最优估计,将该粒子滤波用于目标跟踪,具有较好的过遮挡能力.实验结果表明,该算法状态估计性能好,能够很好地实现复杂图像序列中的目标跟踪.     

基于粒子滤波的红外目标跟踪

粒子滤波(Partic le F ilter)是一种处理非线性和非高斯动态系统状态估计的有效技术.提出了一种基于粒子滤波的红外目标稳健跟踪新方法.在粒子滤波理论框架下,红外目标的状态后验概率分布用加权随机样本集表示,通过这些随机样本的Bayesian迭代进化实现红外目标的跟踪.系统状态转移模型选择为简单的二阶自回归模型,并自适应地确定系统噪声方差.红外目标的描述利用目标区域的灰度分布,该灰度分布通过核概率密度估计建立.通过计算参考目标的灰度分布和目标样本的灰度分布之间的Bhattacharyya距离,建立系统观测概率模型.实验结果表明该方法是有效的和稳健的.     

Structural sparse coding seeds–active appearance model for object tracking

In this paper, we propose a tracking algorithm that can robustly handle appearance variations in tracking process. Our method is based on seeds–active appearance model, which is composed by structural sparse coding. In order to compensate for illumination changes, heavy occlusion and appearance self-updating problem, we proposed a mixture online learning scheme for modeling the target object appearance model. The proposed object tracking scheme involves three stages: training, detection and tracking. In the training stage, an incremental SVM model that directly measures the candidates samples and target difference. The proposed mixture generate–discriminative method can well separate two highly correlated positive candidates images. In the detection stage, the trained weighted vector is used to separate the target object in positive candidates images with respect to the seeds images. In the tracking stage, we employ the particle filter to track the object through an appearance adaptive updating algorithm with seeds–active constrained sparse representation. Based on a set of comprehensive experiments, our algorithm has demonstrated better performance than alternatives reported in the current literature.     

改进高斯混合粒子滤波被动目标跟踪算法

针对被动传感器跟踪系统非线性较强问题,提出了一种基于改进高斯混合粒子滤波的被动传感器目标跟踪算法。该算法基于Sigma点卡曼滤波和粒子滤波的特点,用有限的高斯混合模型来近似后验状态密度、系统噪声和观测噪声的分布。然后结合遗传算法和EM算法来实现模型的降阶,克服了EM算法假定混合成分数为已知、迭代的结果需要依赖初始值、可能收敛到局部最大点或可能收敛到参数空间的边界的缺点,从而改善粒子枯竭的问题。仿真实验结果表明在被动传感器跟踪领域,与传统粒子滤波、基于EM的高斯混合粒子滤波和基于贪心EM的高斯混合粒子滤波相比,该算法在保持高精度估计能力的同时,具有较强的鲁棒性,是解决非线性系统状态估计问题的一种有效方法。     

一种新型多特征融合粒子滤波视觉跟踪算法

针对单一视觉信息在动态变化环境下描述目标不够充分、跟踪目标不够稳定的缺点,提出了一种基于粒子滤波框架的新型多特征融合的视觉跟踪算法。采用颜色和形状信息来描述运动模型,通过民主合成策略将两种信息融合在一起,使得跟踪算法能根据当前跟踪形势自适应调整两种信息的权重以期达到最佳的最大似然比,实现信息间的优势互补。在设计粒子滤波跟踪算法时,利用自适应信息融合策略构建似然模型,提高了粒子滤波跟踪算法在复杂场景下的稳健性。实验结果表明,多特征融合跟踪算法不仅能准确、高效地跟踪目标,而且对光照、姿态变化引起的目标表观变化具有良好的鲁棒性。     

Transferring visual prior for online object tracking

Visual prior from generic real-world images can be learned and transferred for representing objects in a scene. Motivated by this, we propose an algorithm that transfers visual prior learned offline for online object tracking. From a collection of real-world images, we learn an overcomplete dictionary to represent visual prior. The prior knowledge of objects is generic, and the training image set does not necessarily contain any observation of the target object. During the tracking process, the learned visual prior is transferred to construct an object representation by sparse coding and multiscale max pooling. With this representation, a linear classifier is learned online to distinguish the target from the background and to account for the target and background appearance variations over time. Tracking is then carried out within a Bayesian inference framework, in which the learned classifier is used to construct the observation model and a particle filter is used to estimate the tracking result sequentially. Experiments on a variety of challenging sequences with comparisons to several state-of-the-art methods demonstrate that more robust object tracking can be achieved by transferring visual prior.     

基于深度特征表达与学习的视觉跟踪算法研究

该文针对视觉跟踪中运动目标的鲁棒性跟踪问题,将深度学习引入视觉跟踪领域,提出一种基于多层卷积滤波特征的目标跟踪算法。该算法利用分层学习得到的主成分分析(PCA)特征向量,对原始图像进行多层卷积滤波,从而提取出图像更深层次的抽象表达,然后利用巴氏距离进行特征相似度匹配估计,进而结合粒子滤波算法实现目标跟踪。结果表明,这种多层卷积滤波提取到的特征能够更好地表达目标,所提跟踪算法对光照变化、遮挡、异面旋转、摄像机抖动都具有很好的不变性,对平面内旋转也具有一定的不变性,在具有此类特点的视频序列上表现出非常好的鲁棒性。     

Robust observation model for visual tracking in particle filter

A robust observation model for visual tracking is proposed in this paper. The model consists of three appearance models: fixed appearance model, adaptive appearance model, and two-frame appearance model. The three appearance models are used, respectively, for catching unchanged components, slow changes, and rapid changes in object appearance. During tracking, the robust observation model is incorporated in a particle filter, and the particle filter can automatically select proper appearance models to track object according to the current tracking situation. Occlusion analysis is implemented using the M-estimation technique. Experimental results demonstrate that the proposed algorithm can track objects well under many challenging tracking situations.     

非线性滤波算法分析及其性能比较

针对目标跟踪实际应用中量测方程非线性对滤波精度和稳定性的影响,重点分析了模型线性化的滤波算法、无迹卡尔曼滤波（UKF）和粒子滤波算法（PF）的基本原理和特点以及适应的条件。仿真试验比较了扩展卡尔曼、无迹卡尔曼和粒子滤波的跟踪效果,结果表明非线性条件下粒子滤波算法优于其它两种滤波算法。     

基于多特征自适应融合的分类采样跟踪算法

针对目标跟踪中的场景易变和目标模板不稳定等问 题,提出了一种基于多特征自适应 融合的分类采样跟踪算法。算法利用密集特征信息将目标模板用多个重叠子区域划分,每 个子区域对应一个多特征采样窗口。利用多特征自适应融合构造强可区分性的目标模型,最 大程度地提高各子区域之间的互补性,以增强目标模板的区分能力。在粒子滤波(PF)框架下 , 多特征自适应融合策略提高了目标观测质量,保证跟踪的持续稳定。实验结果表明,本文所 提算法具 有良好的目标跟踪性能,并对动态场景、目标形变及遮挡情况具有较好的跟踪准确性和鲁棒 性。     

基于时序特性的自适应增量主成分分析的视觉跟踪

当前基于增量主成分分析(PCA)学习的跟踪方法存在两个问题,首先,观测模型没有考虑目标外观变化的连续性;其次,当目标外观的低维流行分布为非线性结构时,基于固定频率更新模型的增量PCA学习不能适应子空间模型的变化。为此,该文首先基于目标外观变化的连续性,在子空间模型中提出更合理的目标先验概率分布假设。然后,根据当前跟踪结果与子空间模型之间的匹配程度,自适应调整遗忘比例因子,使得子空间模型更能适应目标外观变化。实验结果验证了所提方法能有效提高跟踪的鲁棒性和精度。     

一种易于初始化的类卷积神经网络视觉跟踪算法

该文针对视觉跟踪中运动目标的鲁棒性跟踪问题,基于主成分分析(PCA)和卷积神经网络(CNN),提出一种易于初始化的类CNN提取深度特征的视觉跟踪算法。该算法首先利用仿射变换对原始图像进行处理,然后对归一化尺寸的图像进行分层PCA学习,将学习得到的PCA特征向量作为CNN结构中的各阶滤波器,完成特征提取网络的初始化,再利用特征提取网络获取目标的深层次表达。最后结合粒子滤波,利用一个简单的逻辑回归分类器通过分类估计实现目标跟踪。结果表明,利用这种易于初始化的CNN提取到的深度特征能够有效地区分目标和背景,具有很好的可区分性,提出的视觉跟踪算法对光照变化、尺度变化、遮挡、旋转和摄像机抖动等都具有良好的适应性,在许多视频序列上表现出了较好的鲁棒性和准确性。     

基于显著性检测的星载嵌入式跟踪系统

随着航天科技的不断发展,计算机视觉算法在卫星上的应用方兴未艾,为了实现更多的功能需求和应对可能的威胁,视觉目标跟踪作为其中基础但具有挑战性的任务更是至关重要。然而,目前已有的目标跟踪算法大多数算法只限于对图像序列进行跟踪。另一方面,受到硬件条件制约,很多优秀的算法因为复杂度较高很少被应用到星载嵌入式系统中。这些目标跟踪算法运行时,通常需要人为地给出目标的边界框。为了自动得到边界框,需要目标检测算法对输入图像进行运动目标检测。本文提出了一种基于显著性检测和相关滤波的单目标检测与跟踪一体化算法,并与嵌入式系统相结合,在搭载的TMS320C6678芯片上达到了2 048 pixel×2 048 pixel分辨率下24 fps的帧率。具体地,检测算法负责对图像进行预处理并获得边界框,然后目标跟踪算法给出目标在后续帧中的位置。为了验证算法在实际跟踪中的有效性,本研究搭建了一个由相机、DSP和云台组成的光学平台并进行了实验验证。在该系统中,DSP自动完成检测、跟踪、驱动云台和再检测任务,达到了很好的检测跟踪效果。     

一种基于均值移位的红外目标跟踪新方法

均值移位算法是一种搜索与样本点分布最相近模式的非参数统计方法．在彩色序列图像目标跟踪中,均值移位算法是一种有效的方法．但在红外目标跟踪中,由于单一灰度特征空间缺乏描述红外目标的信息,使得基于均值移位算法的红外目标跟踪不稳健．为了克服这个缺点,提出了构造级联灰度空间的红外目标跟踪新方案．同时,对于不同的红外图像序列使用不同的方法产生级联灰度空间．实验结果表明该方法对于红外小目标以及强杂波背景目标的跟踪是有效和稳健的．     

基于激光点云坐标系的3D多目标跟踪算法研究

3D多目标跟踪是无人车领域中一个核心技术之一。3D多目标跟踪的相关工作多通过设计复杂的目标建模算法或数据关联算法,以寻求提高3D多目标跟踪系统的鲁棒性。为降低3D多目标跟踪系统的复杂性,将目标的三维中心点视为该目标进行跟踪,提出了基于激光点云坐标系的3D多目标跟踪算法。首先,使用3D目标检测器检测出每帧激光点云中的目标。然后,用卡尔曼滤波器预测上一帧目标的三维中心点在当前帧的位置状态,融合激光雷达坐标系下相邻两帧之间目标三维体积的交并比与目标三维坐标中心点坐标之间的欧式距离作为度量尺度,使用贪婪算法匹配最近邻的目标。在KITTI跟踪数据集上的实验结果表明,所提出的多目标跟踪算法表现优异,运行速度达到了63 f/s,且车辆类的sAMOTA达到了9432。     

Face detection and tracking using a Boosted Adaptive Particle Filter

A novel algorithm, termed a Boosted Adaptive Particle Filter (BAPF), for integrated face detection and face tracking is proposed. The proposed algorithm is based on the synthesis of an adaptive particle filtering algorithm and the AdaBoost face detection algorithm. An Adaptive Particle Filter (APF), based on a new sampling technique, is proposed. The APF is shown to yield more accurate estimates of the proposal distribution and the posterior distribution than the standard Particle Filter thus enabling more accurate tracking in video sequences. In the proposed BAPF algorithm, the AdaBoost algorithm is used to detect faces in input image frames, whereas the APF algorithm is designed to track faces in video sequences. The proposed BAPF algorithm is employed for face detection, face verification, and face tracking in video sequences. Experimental results show that the proposed BAPF algorithm provides a means for robust face detection and accurate face tracking under various tracking scenarios.     

基于颜色的粒子滤波算法的改进与全硬件实现

传统基于颜色的粒子滤波算法在硬件实现中存在着跟踪效果不理想、实时性差等问题.该文结合硬件电路需要对基于颜色的粒子滤波算法进行了改进,在传统SR重采样算法的基础上将剩余粒子撒向目标点附近,以提高其在硬件环境下跟踪的准确性与稳定性.文中给出了改进算法的全硬件实现的电路架构,并在FPGA上完成了目标跟踪系统的实现.实验表明提...     

基于iOS设备的自主航拍巡检控制终端的设计

针对无人机航拍巡检操作繁琐、效率低下等问题,提出了一种基于iOS设备和图像跟踪算法的无人机自主航拍巡检控制终端.使用无人机控制软件开发包搭建了终端与无人机、航拍云台、无线图像传输设备、无人机控制器之间的通信链路.基于FFmpeg实现视频解码,提出了基于自适应尺度的时空上下文跟踪算法,引入尺度参数变化率以在线纠正更新尺度参数的过激变化,大幅增强了航拍云台在复杂背景下的跟踪稳定性,实现了无人机航拍系统对不同尺度目标的自主、稳定跟踪,提高了航拍巡检效率.