基于形态匹配聚类的近邻扩展目标跟踪算法

针对传统扩展目标跟踪（Extended Target Tracking, ETT）算法在处理近邻目标时面临的计算效率低下和跟踪不准确的问题,提出了一种形态匹配聚类量测集划分与高斯逆威沙特概率假设密度（Gaussian Inverse Wishart Probability Hypothesis Density, GIW-PHD）滤波器相结合的跟踪处理方法。该方法首先由GIW-PHD滤波器得到预测的目标状态,其次使用DBSCAN(Density-Based Spatial Clustering of Applications with Noise, DBSCAN)算法完成量测集的初步划分,在此基础上利用较高权重的预测分量实现对多个近邻目标混合量测簇的判断,进而使用椭圆形状约束（Elliptic Shape Constraint, ESC）的FCM(Fuzzy C-Means, FCM)算法（ESC-FCM）对混合簇进行二次划分得到更精确的划分结果,最后将划分结果合并后送入GIW-PHD滤波器完成目标状态的更新。仿真结果表明,本文所提量测集划分方法能够充分利用GIW-PHD滤波器预测步获取...     

一种基于多模型高斯逆Wishart PHD滤波器的空间邻近目标跟踪方法

将空间邻近目标(Closely Spaced Objects,CSOs)整体建模为扩展目标(Extended Target,ET),用随机矢量和随机矩阵分别描述CSOs质心运动和扩散状态,并采用高斯逆Wishart(Gaussian inverse Wishart,GIW)概率假设密度(Probability Hypothesis Density,PHD)滤波器实现杂波和漏检条件下CSOs的稳定跟踪.修正了原GIW-PHD滤波器量测模型和形状估计的缺陷,给出新的递推表达式,并在此基础上提出一种多(形变)模型GIW-PHD滤波器,以适应CSOs分裂和融合引起的形状变化.仿真结果表明,所提算法能够有效跟踪CSOs,状态估计比原GIW-PHD更加准确,对CSOs的变化更加敏感.     

一种基于粒子滤波的双极化雷达检测前跟踪算法

针对雷达在低信噪比（signal to noise ratio,SNR）条件下对运动目标的检测和跟踪难题,提出了一种基于粒子滤波（particle filter,PF）的双极化雷达运动目标检测前跟踪（track before detect,TBD）算法,又称联合粒子滤波检测前跟踪（joint particle filter-track before detect,JPF-TBD）方法.该算法借鉴传统的TBD算法处理框架,以经典PF算法为基础,使用双通道幅度相位似然比函数计算粒子权值,并实现了完整的PF过程.与同类研究相比,所提算法能够充分利用双极化雷达各通道幅度和相位信息,进一步扩展了PF算法的应用范围.仿真实验表明:在SNR>10 dB,虚警概率为10-6的情况下所提算法对目标的检测概率大于0.8.     

基于相关滤波扩展记忆系统模型的智能目标跟踪

将目标跟踪过程看作一个多重记忆系统模型,提 出了基于相关滤波的扩展记忆系统模型,实现了基 于记忆系统模型的智能目标跟踪。首先,通过提取跟踪目标特征学习目标信息,生成短时相 关滤波器,产 生短时记忆;然后利用每一帧短期记忆的不断重复与更新,产生长时记忆,生成长时相关滤 波器。短时与 长时记忆构成相关滤波记忆系统模型,完成目标跟踪。在此模型基础上,分析与挖掘模型中 的相关滤波数 据,加入四种智能化控制信息,构建扩展记忆系统模型,实现智能化的目标跟踪。基于相关 滤波的扩展记 忆系统模型利用生物记忆的原理使目标跟踪更加自动化、智能化,增强目标跟踪的准确性。 实验结果表明, 与当前流行的相关滤波跟踪算法相比,本文算法提高了目标跟踪的抗干扰性、抗遮挡性与抗 形变能力,同时保证了在尺度跟踪的有效性。     

基于PPP模型的多扩展目标跟踪的JPDA算法研究

针对多扩展目标跟踪问题,提出了基于泊松点过程（Poisson Point Process, PPP）模型的多扩展目标跟踪的联合概率数据关联（Joint Probabilistic Data Association, JPDA）算法。首先,采用PPP对扩展目标进行测量建模,其次以“多对一”关联模型思想提出一种的JPDA算法,从而计算运动目标的当前有效量测的边缘关联概率,然后结合该边缘关联概率以概率数据关联（Probability Data Association, PDA）的方式分别更新每个扩展目标的运动参数和形状参数向量,最后通过仿真实现了当扩展目标相互靠近或出现交叉时的跟踪。实验结果表明,在高杂波环境下,本文所提出的算法在计算时间和跟踪稳定上具有较明显的优势。      

单站无源定位原理浅析

探讨了一种基于目标频域和空域参数测量信息，利用固定单站对机动目标进行无源定位与跟踪的算法，并详细分析了定位原理。在建立目标机动模型与测量方程的基础上，运用修正增益的扩展卡尔曼滤波（MGEKF）算法，实现对机动目标进行定位与跟踪。     

IMM迭代扩展卡尔曼粒子滤波跟踪算法

该文提出了一种交互式多模型(IMM)迭代扩展卡尔曼粒子滤波机动目标跟踪算法。该算法在多模型中使用了改进的粒子滤波器,通过对迭代扩展卡尔曼滤波(IEKF)的测量更新按照高斯牛顿方法进行修正,减小了非线性滤波带来的线性化误差,然后利用修正的IEKF来产生粒子滤波的重要性密度函数,使其融入最新观测信息。最后将所提算法与交互式多模型粒子滤波(IMMPF)进行了比较,仿真结果表明该算法具有更好的跟踪性能。     

一种适用于分布式摄像机网络的SCIWCF算法

本文使用容积卡尔曼滤波器来处理分布式摄像机网络中的目标跟踪问题.平方根容积信息滤波（Square-Root Cubature Information Filter,SCIF）是容积卡尔曼滤波的一种扩展,其具有有效性和可靠性等方面优势,有利于对多源信息进行信息融合.然而当该算法应用于像摄像机网络这种大规模网络时,如果采用一般的集中式处理,中心节点可能会承受较大的计算压力.针对这个问题,本文首先将平方根容积信息滤波器进行了扩展,提出分布式平方根容积信息滤波器,使其能适应大规模网络.另外在摄像机网络中,由于摄像机装置在一个较大的区域内,由于摄像机观测区域有限,目标可能会出现在观察的盲区,这样就会存在某些摄像机的测量数据无效.针对这个问题,本文提出了平方根容积信息加权一致性滤波器（Square-Root Cubature Information Weighted Consensus Filter,SCIWCF）对状态信息和信息矩阵加权,减小这些无效信息在一致性算法的作用,从而提高整体的滤波性能.仿真实验结果表明,本文提出的算法能够在摄像机网络中对目标进行有效跟踪,在估计精度和滤波器稳定性等方面要优于传统的信息滤波.     

一种改进的高斯逆威沙特概率假设密度扩展目标跟踪算法

假设扩展目标(ET)的扩展和量测数目分别为椭圆和泊松模型,高斯逆威沙特概率假设密度(GIW-PHD)能够估计扩展目标的运动和扩展状态。然而,该滤波器对空间邻近目标的数目、非椭圆目标和受到遮挡目标的扩展估计不够准确。针对这些问题,该文提出一种改进的GIW-PHD。首先,假设目标扩展为一个相同尺寸的参考椭圆,通过设计新的散射矩阵得到改进的随机矩阵(RM)方法。然后,将改进的RM方法与假设量测数目服从多伯努利分布的ET-PHD结合,得到改进的GIW-PHD滤波器。仿真和实验结果表明,与传统GIW-PHD相比,改进的GIW- PHD估计的目标数目和量测数目较多,扩展较大的椭圆和非椭圆目标的扩展更准确。     

机动目标无源跟踪IMM-MKF算法

为提高运动多站对机动目标的无源跟踪性能,提出了一种新的基于交互式多模型-边缘化卡尔曼滤波(IMM-MKF)的机动目标跟踪算法。该算法将交互式多模型(IMM)结构和边缘化卡尔曼滤波(MKF)结合,利用MKF算法对每个模型进行滤波,对滤波结果进行交互作用来得到跟踪结果。以只测角机动目标跟踪为例对所提算法进行仿真分析,仿真结果表明,相对于采用扩展卡尔曼滤波(EKF)、不敏卡尔曼滤波(UKF)及容积卡尔曼滤波(CKF)算法的典型交互式多模型算法,所提算法具有更好的跟踪性能。     

结合幅度信息的扩展目标随机有限集跟踪方法

基于随机有限集的扩展目标跟踪方法通常根据量测的空间信息进行量测划分,在杂波密集环境下有可能将杂波量测划入目标单元,从而造成跟踪性能的下降。为此,该文将目标和杂波的幅度信息引入高斯逆威沙特概率假设密度(GIW-PHD)滤波器,通过计算量测子集的幅度似然寻找最优的量测划分方法。此外,计算量测单元的中心时,采用幅度加权的方法计算量测单元的质量中心,以取代目前广泛使用的几何中心,从而进一步降低杂波对滤波器的干扰。在信杂比分别为13 dB和6 dB的条件下,通过对Rayleigh杂波中Swerling 1型起伏目标的跟踪结果证明了所提方法相比高斯逆威沙特概率假设密度滤波器具有更优的势估计和状态估计性能。      

Efficient and robust fragments-based multiple kernels tracking

Representing an object with multiple image fragments or patches for target tracking in a video has proved to be able to maintain the spatial information. The major challenges in visual tracking are effectiveness and robustness. In this paper, we propose an efficient and robust fragments-based multiple kernels tracking algorithm. Fusing the log-likelihood ratio image and morphological operation divides the object into some fragments, which can maintain the spatial information. By assigning each fragment to different weight, more robust target and candidate models are built. Applying adaptive scale selection and updating schema for the target model and the weighting factors of each fragment can improve tracking robustness. Upon these advantages, the novel tracking algorithm can provide more accurate performance and can be directly extended to a multiple object tracking system.     

无线传感器网络的目标跟踪算法

为了解决无线传感器网络跟踪非线性运动目标的分布式数据融合问题,使用了基于扩展信息滤波器(EIF)的分布式估计算法.对于活跃传感器的选择方法,采用了基于与目标位置接近程度的近邻选择算法和基于信息贡献的信息选择算法.仿真结果表明,与分布式扩展信息滤波器(DEIF)算法相比,近邻选择算法和信息选择算法得到了相似的响应曲线,且具有减少能量消耗和简化计算的优点.     

Object Tracking via Partial Least Squares Analysis

We propose an object tracking algorithm that learns a set of appearance models for adaptive discriminative object representation. In this paper, object tracking is posed as a binary classification problem in which the correlation of object appearance and class labels from foreground and background is modeled by partial least squares (PLS) analysis, for generating a low-dimensional discriminative feature subspace. As object appearance is temporally correlated and likely to repeat over time, we learn and adapt multiple appearance models with PLS analysis for robust tracking. The proposed algorithm exploits both the ground truth appearance information of the target labeled in the first frame and the image observations obtained online, thereby alleviating the tracking drift problem caused by model update. Experiments on numerous challenging sequences and comparisons to state-of-the-art methods demonstrate favorable performance of the proposed tracking algorithm.     

Direct tracking of noncircular sources for multiple arrays via improved unscented particle filter method

Direct tracking problem of moving noncircular sources for multiple arrays is investigated in this study. Here, we propose an improved unscented particle filter (I-UPF) direct tracking method, which combines system proportional symmetry unscented particle filter and Markov Chain Monte Carlo (MCMC) algorithm. Noncircular sources can extend the dimension of sources matrix, and the direct tracking accuracy is improved. This method uses multiple arrays to receive sources. Firstly, set up a direct tracking model through consecutive time and Doppler information. Subsequently, based on the improved unscented particle filter algorithm, the proposed tracking model is to improve the direct tracking accuracy and reduce computational complexity. Simulation results show that the proposed improved unscented particle filter algorithm for noncircular sources has enhanced tracking accuracy than Markov Chain Monte Carlo unscented particle filter algorithm, Markov Chain Monte Carlo extended Kalman particle filter, and two-step tracking method.     

基于颜色与深度信息特征融合的一种多目标跟踪新算法

提出了一种基于颜色与深度信息特征融合的 可逆跳的马尔科夫链蒙特长洛(RJMCMC)多目标跟踪算法。首先,融合彩色信息和深度信息对 运动目标进行检测;然后,根据多目标检测的结果建 立观测似然模型,并构建合理的状态转移模型;最后,通过RJMCMC粒子滤波算法实现多目 标跟踪。实验 结果表明,本文提出的多目标跟踪算法具有较强的鲁棒性,能够稳定的跟踪多目标,具有较 高的准确率。     

Active contours for video object tracking using region, boundary and shape information

In this paper, we propose a robust model for tracking in video sequences with non-static backgrounds. The object boundaries are tracked on each frame of the sequence by minimizing an energy functional that combines region, boundary and shape information. The region information is formulated by minimizing the symmetric Kullback–Leibler (KL) distance between the local and global statistics of the objects versus the background. The boundary information is formulated using a color and texture edge map of the video frames. The shape information is calculated adaptively to the dynamic of the moving objects and permits tracking that is robust to background distractions and occlusions. Minimization of the energy functional is implemented using the level set method. We show the effectiveness of the approach for object tracking in color, infrared (IR), and fused color-infrared sequences.     

基于扩展空间直方图的红外目标跟踪方法

针对传统均值漂移算法中仅仅利用目标的颜色信息而导致目标模型分辨能力不高的问题,提出了一种基于扩展空间直方图的红外目标均值漂移跟踪方法.首先对空间直方图进行扩展,构建了一种结合目标颜色分布和空间约束关系的联合空间颜色模型,有效提高了目标模型的分辨能力.通过给定目标空间位置和颜色联合概率密度函数,定义目标区域与候选区域概率密度的相似性度量,进而实现了红外目标的准确定位.实验结果表明该算法简单有效,能准确跟踪前视红外目标.     

Robust online tracking via sparse gradient convolution networks

Convolution networks trained offline have recently exhibited promising performance in object tracking tasks. However, offline training is time-consuming and their performance heavily rely on the category of auxiliary training sets. In this paper, we propose a sparse gradient convolution network without pretraining for object tracking. This approach combines shallow convolutional networks and traditional methods (gradient features and sparse representations) to avoid the offline training. In the first frame, we utilize the sparse representation method to learn a series of gradient-based local patches served as fixed filters, and they are used to convolving the input image in the subsequent frames to encode local structural information. Then, we stack all the local structure features to construct global spatial structure features, and the inner geometric layout information is preserved. Moreover, sparse coding and online updating are used to overcome issues related to target appearance variations. Qualitative and quantitative evaluations based on a challenging benchmark dataset demonstrate the effectiveness of the proposed algorithm against several state-of-the-art tracking methods.